Just Say Sepsis! A review of the process of care received by patients with sepsis

Improving the quality of healthcare Just Say Sepsis! A review of the process of care received by patients with sepsis

A report by the National Confidential Enquiry into Patient Outcome and Death (2015)

Compiled by: APL Goodwin FRCA FFICM – Clinical Co-ordinator Royal United Hospital Bath NHS Trust

V Srivastava FRCP (Glasg) MD – Clinical Co-ordinator King’s College Hospital NHS Foundation Trust

H Shotton PhD – Clinical Researcher

K Protopapa BSc Psy (Hons) – Researcher

A Butt BSc (Hons) – Research Assistant

M Mason PhD – Chief Executive

The study was proposed by: UK Sepsis Trust – Dr Ron Daniels and Public Health England – Dr Imogen Stephens

The study was commissioned by the Healthcare Quality Improvement Partnership (HQIP) on behalf of NHS England, NHS Wales, the Northern Ireland Department of Health, Social Services and Public Safety (DHSSPS), the States of Guernsey, the States of Jersey and the Isle of Man government.

The authors and Trustees of NCEPOD thank the NCEPOD staff for their work in collecting and analysing the data for this study: Robert Alleway, Donna Ellis, Heather Freeth, Dolores Jarman, Kathryn Kelly, Kirsty MacLean Steel, Nicholas Mahoney, Eva Nwosu, Neil Smith and Anisa Warsame. Contents

Acknowledgements 3 Foreword 5 Principal recommendations 9 Introduction 11 1 – Method and data returns 13 2 – Organisational data 17 Key findings 33 3 – Patient population and pre-hospital care 35 Case study 1 47 Case study 2 48 Case study 3 50 Key findings 55 4 – Admission to hospital 57 Case study 4 60 Key findings 61 5 – Patients with hospital-acquired infections 63 Key findings 66 6 – First identification of sepsis 67 Case study 5 70 Case study 6 72 Case study 7 76 Key findings 76 7 – Initial management of sepsis 77 Case study 8 87 Case study 9 92 Key findings 94 8 – Complications of sepsis and discharge planning 95 Key findings 104 9 – Overall quality of care 105 Recommendations 107 Summary 109 References 111 Appendices 115 1 - Glossary 115 2 - Existing sepsis information, templates and tools 117 3 - The role and structure of NCEPOD 118 4 - Participation 120 Acknowledgements Back to contents

This report, published by NCEPOD, could not have been Dr Chris Custard, Consultant in Acute & General achieved without the support of a wide range of individuals Internal Medicine who have contributed to this study. Ms Beryl De Souza, Consultant Plastic Surgeon Dr Paul Dean, Consultant in Anaesthesia and Intensive Our particular thanks go to: Care Medicine Mrs Mary Edwards, Critical Care Outreach Team Lead The Study Advisory Group who advised NCEPOD Dr Rhian Edwards, Consultant in Anaesthesia and on the design of the study: Intensive Care Medicine Ms Susan Bracefield, Critical Care Nurse Ms Karin Gerber, Clinical Nurse Specialist, Critical Professor Gordon Carlson, Consultant Surgeon Care Outreach Dr Niall Collum, Consultant in Emergency Medicine Ms Carmel Gordon, Critical Care Outreach Lead Dr Alaric Colville, Consultant in Microbiology Ms Patricia Guilfoyle, Critical Care Outreach Sister Professor Matthew Cooke, Consultant in Ms Rhona Hayden, Matron Critical Care Outreach Services Emergency Medicine Lt Col Paul Hunt, Military Consultant in Emergency Medicine Dr Ron Daniels, Consultant Intensivist Dr Stephen Hutchinson, Consultant in Anaesthesia and Mr Chris Hancock, Programme Manager, NHS Wales Intensive Care Medicine Sepsis Programme Dr Phil Jacobs, Consultant Physician in Acute and Mr Eddie Morris, Consultant Obstetrician and Gynaecologist General Medicine Professor Sebastian Lucas, Consultant Pathologist Mr Sunjay Kanwar, Consultant General and Upper Mr Dermot O’Riordan, Consultant Surgeon Gastrointestinal Surgeon Dr Marilyn Plant, General Practitioner Dr Liza Keating, Consultant Intensive Care and Dr Christopher Roseveare, Consultant Physician Emergency Medicine Dr Andrew Stacey, Consultant in Microbiology Dr Jeff Keep, Consultant in Emergency Medicine Dr Imogen Stephens, Consultant in Public Health Medicine & Major Trauma Ms Catherine White, Patient Representative, ICUSteps Dr Alastair Keith, Clinical Fellow in Anaesthesia Dr Ian Kerslake, Consultant in Anaesthesia and Intensive The Reviewers who peer reviewed the cases: Care Medicine Mr John Abercrombie, Consultant General Surgeon Dr Patrick Lillie, Consultant in Acute Medicine and Dr Julie Andrews, Consultant in Medical Microbiology Infectious Diseases and Virology Dr Stephen Luney, Consultant Neuroanaesthetist Ms Esme Blyth, Sepsis Nurse and Critical Care Dr Cliff Mann, Consultant in Emergency Medicine Outreach Nurse Dr Chakri Molugu, Consultant in Acute and General Dr Caroline Burford, Senior Trainee in Acute and Intensive Medicine Care Medicine Dr Nick Murch, Consultant in Acute Medicine Dr John Bye, General Practitioner Mr Julian Newell, Emergency Nurse Practitioner Dr Tim Collins, Acute Care Consultant Nurse Dr Emmanuel Nsutebu, Consultant in Infectious Diseases Dr Ben Creagh-Brown, Consultant Physician in Intensive and General Internal Medicine Care and Respiratory Medicine Dr Tim Nutbeam, Consultant in Emergency Medicine Ms Nicola Credland, Specialist Lecturer Practitioner Dr Stuart Nuttall, Consultant in Emergency Medicine

3 Acknowledgements

Dr Helena Parsons, Consultant in Microbiology Ms Catherine Plowright, Consultant Nurse in Critical Care Dr Lee Poole, Consultant in Anaesthesia and Intensive Care Medicine Mr Suman Shrestha, Advanced Critical Care Nurse Practitioner Dr Hannah Skene, Consultant in Acute and General Medicine Dr Catherine Snelson, Consultant in Critical Care and Acute Medicine Dr Mike Spivey, Consultant in Intensive Care and Anaesthesia Dr Shiva Sreenivasan, Consultant in Acute and General Medicine Dr Jonathan Stacey, General Practitioner Dr Pius Tansinda, Consultant in Nephrology Dr Jerry Thomas, Consultant in Anaesthesia and Intensive Care Medicine Dr Madhankumar Vijaya Kumar, Consultant in Anaesthesia and Intensive Care Medicine Dr Rosanne Wrench, General Practitioner Mr Martin Wiese, Consultant Emergency Physician

Thanks also go to all the NCEPOD Local Reporters, NCEPOD Ambassadors, Study Specific Contacts and the Clinicians who took the time to complete questionnaires.

4 Foreword Back to contents

This NCEPOD study addresses a huge subject, which sets also tell us that in a significant number of cases the culprit it aside from those of our reports that have focused on now turns out to be viral or fungal, although that did not Cinderella topics, parts of the NHS that have been previously seem to be so in the cases we studied. overlooked. Sepsis, by which we mean the systemic inflammatory response to microbial infection, causing The laity are well aware that medicine is constantly evolving damage to organs then shock and ultimately death is a as new interventions are introduced. To discover that common problem: the international prevalence is estimated1 the bugs are doing the same takes a bit of getting used at 300 per 100,000, suggesting that there are around to: these are not necessarily bacteria that are developing 200,000 cases a year in the UK alone. To put this into new resistances or sharing their resistances to individual the context of our recent studies, there are around 5,500 antibiotics through species-jumping, although no doubt lower limb amputations, a similar number of subarachnoid that is adding to the problem. Here the nature of the haemorrhages, 8,000 or so aortic aneurysms, 9,000 deaths disease is becoming both more complicated and more from alcohol-related liver disease, 10,000 tracheostomies elusive. Critics must acknowledge that medicine has never and 11,600 bariatric procedures performed per year. Sepsis been easy and here it is getting more difficult. eclipses even the 90,000 patients treated for gastrointestinal bleeding. Finally, this is a significant study precisely because the importance of the issue has been recognised by others, Sepsis is important because it is a major cause of avoidable including ongoing work at NHS England and the ‘1000 Lives death in our hospitals. According to the same source1, Plus’ national improvement programme in Wales. The role the current mortality from sepsis is greater than that from of this report is not to draw attention to an unrecognised myocardial infarction in the 1960’s. We are told that the problem but to examine an acknowledged problem much survival from sepsis-induced hypotension is over 75% if it more closely than before and to demonstrate how these is recognised promptly, but that every hour’s delay causes patients are being let down. Here you will learn the nature of that figure to fall by over 7%,2 implying that the mortality the malady and the remedies our peer reviewers prescribe. increases by about 30%. Since we found that longer delays in treatment are commonplace, the results of this study The first thing that strikes me is that being a big subject should make everyone sit up and take notice. has enormous advantages. Since the problem is so common, the NHS can justify providing appropriate Sepsis is also important because it has become more resources to cater for it. That is not as easy as it sounds difficult to manage. It used to be viewed as a disease because it also commonly presents in the community. invariably caused by gram-negative bacteria, but the picture Nevertheless it is straightforward to provide the protocols has changed over the last 30 years. In over 60% of the cases and resources needed in a network across the nation. If it we reviewed, no pathogen was ever identified and where happens well over 100,000 times a year and the patients bacteria were implicated, the majority were gram-positive remain in hospital for some time, there will at least be no bacteria that either refused to be cultured or to disclose shortage of cases in most of our acute hospitals, so there their sensitivities when they are cultured made up a majority should be plenty of opportunities to train clinicians to of the pathogens these physicians had to treat. Our experts recognise and deal with the condition.

1 National Clinical Effectiveness Committee of Ireland – Nov 2014 No 6 - Sepsis 2 See page 81

5 Foreword

It should also be relatively easy to put the armamentarium The reasons why poor practice is occurring more often at the disposal of the clinicians to enable them to manage than we would hope are varied but they will not come the initial interventions effectively. I was struck by how very as any surprise to readers of recent NCEPOD reports, simple are the components of good first line treatment, because at root the themes are familiar. Sick patients which in many cases will hold the key to optimising survival. have to be identified by their families and carers, then The Sepsis Six is a well-known set of 3 investigations and assessed by clinicians and nurses who are able to 3 initial therapies, none of which are remotely surprising: recognise the diagnosis because they have seen it before. start the patient on high flow oxygen and take blood for In the Emergency Departments cases of sepsis have to culture before treatment starts: put up intravenous fluids be distinguished from the other 99% of the 24 million and give intravenous antimicrobials and later change them new cases now coming to our hospitals every year.3 Those if the cultures suggest they are wrong; measure the lactate, cases that occur amongst existing inpatients may be even do a full blood count, and monitor the urine output. None harder to spot because the onset may be insidious and the of these first line interventions are beyond the competence assessment of the problem may be less analytical on the of juniors or the resources of the departments in which part of those who are managing an existing condition. they work. Of course the management of the later onset of multi organ failure is extraordinarily difficult and will usually One broad underlying problem is that the recognition of require sub-specialist involvement, as will the management illness often requires more experience than junior members of culture-negative organisms, but a good outcome for of staff can draw upon. To mitigate this problem, first many appears to depend primarily on recognising the NCEPOD and subsequently the Royal Colleges and NHS problem and doing the simple things right and promptly. England have called for all acute admissions to be seen by a consultant within 12 or 14 hours. Although welcome Sadly that is not what is happening in the majority of our as an initiative to improve patient care, this will inevitably cases: about one third received good care in the opinion diminish the importance of the assessment by the juniors of our peer reviewers, who are mainstream healthcare and the experience they are acquiring, which will need to be professionals treating this condition on a regular basis. factored into future training. They asked themselves a simple question: is this a standard of care that I would accept from my own team, or is However according to this report neither the subtlety of there room for improvement? These are not harsh or the disease nor the inexperience of the doctors seems to be exacting standards posed by a critic trying to score a point. the mainspring of the problem. It may be very difficult to NCEPOD takes enormous care to ensure that our reports make the diagnosis of sepsis in some cases, but if you are are not skewed, either by unrealistic enthusiasts for a cause confronted with a suspected infection and do not measure or by dyed-in-the-wool defenders of a status quo that the vital signs you are not enhancing your chances. This is the they expect to be mediocre. Having agreed on what they first study where we have looked at events in general practice are looking for before they start, our peer reviewers have and we must be tentative in advising colleagues: here we to explain and defend their conclusions to the group of only had three GP peer reviewers looking at a handful of sets colleagues with whom they are studying the cases in our of notes: I know from my own work that GPs still vary more offices in every case. than hospital doctors in the ways in which they record events. There were 129 cases and in 52 of them our peer reviewers We believe that by this process we articulate the views of sensibly decided that there was too little data for them to be the mainstream professions treating this disease on a daily able to say whether the diagnosis had been missed. basis. The value of the vignettes is that they spell out to the rest of us what the peer reviewers mean when they say We can all understand that it may be difficult for a GP something is good practice or leaves room for improvement to know whether to give an antimicrobial intravenously, in terms that all of us can understand. especially if they do not carry the equipment to take

3 NHS Health and Social Care Information Centre

6 blood for culture at the same time. But in one third of the of severe sepsis. Given that delay may increase the mortality cases we reviewed, not one of the four basic vital signs of by 30% if the patient has progressed to hypotension, it is a temperature, pulse, blood pressure and respiratory rate had “golden hour” for these sepsis patients just as much as it is been recorded.4 When patients were sent in to hospital no for the patient with a coronary thrombosis or stroke. Today referral letter was available in 43% of cases.5 we have a national network of primary angioplasty centres served by teams in which paramedics have been trained As I say, our authors and peer reviewers are well aware that to play a key part, all dedicated to the proposition that they have to get used to a new environment and to develop every minute lost represents the loss of cardiac muscle. The an understanding of the variety of different ways in which same sort of progress is being made in stroke services, with GPs work, but it looks to me as if NCEPOD has found a Hyperacute Stroke Units being supported in the same ways rich field in which they can contribute the insights of their so that appropriate therapies can be delivered within very colleagues to GPs up and down the country. narrow windows of opportunity. Despite the high profile attention, it seems that infection still does not get the same Furthermore, when these patients reached the Emergency priority in practice. Department things did not always improve very much. The number of measurements went up, but a full set of vital This study is unusual in that our peer reviewers found far signs were recorded in only 40% of cases. For the most part less to criticise in the organisation of care than they did in the diagnosis was not made when it should have been and the clinical delivery of that care. This reflects the fact that when it was made, the presumed source of infection was not the organisation can be so much simpler than in other recorded. In a significant minority of cases, no proper record diseases. The administration of an antibiotic and fluids of obligatory first line investigations was made. Of course this intravenously takes a fraction of the resources needed to may reflect that they had been noted to be normal and the maintain a primary angioplasty service or a Hyperacute clinicians caring for them were too busy to make a written Stroke Unit. If the treatment can be delivered at every record of normal findings. Unfortunately, as I so often have to Emergency Department, there is less need to train the tell my clients, the default position is that if it was not written paramedics to make the diagnosis so as to take their down, broadly speaking it did not happen. Furthermore, in patients to the optimal location. However, one of the this disease more than most, a change in the vital signs can most disturbing tables in this remarkable report concerns be a sensitive pointer to a change in the patient’s condition, the baseline therapies started by the hospitals before they however busy the doctor is at the time. transferred patients, having presumably recognised that they needed specialist care.6 Knowing that the patient is Given that this is a progressive disease it is not surprising being moved because they are acutely ill and will often that the clinicians found it easier to recognise that their take several golden hours before they have been settled patients were ill when their condition had evolved and the and assessed by another doctor it seems obvious that they severity had become more obvious. But that is precisely why should be started on an antimicrobial (which entails that the protocols and pathways are designed to elicit the subtle blood samples must first be taken for culture) and given gradations that may reveal a downward trend in time to IV fluids and oxygen to support them during the transfer make the difference between life and death. They are there period when there may be no doctor with them. How to help the inexperienced who understandably find it harder could that not happen in the majority of cases? It is hard to to recognise the sick patient, but the evidence seems to be escape the conclusion that an appropriate sense of urgency that they are not being used consistently. is lacking in far too many cases.

Treatment seems to be no more disciplined than diagnosis. This is the 13th and last report I will have the privilege of I was astonished to read that the majority of patients do not presenting. I have been working with NCEPOD for 11 years, receive antimicrobial drugs within an hour of the diagnosis for the last 6 as its Chair and it is time to hand over. I do

4 See Table 3.22 5 Table 3.23 6 Table 2.19 7 Foreword

not know how the organisation will change in the future, co-ordinators who supervised them; then the research team but if the past is anything to go by, NCEPOD will continue at NCEPOD who collated and analysed the data, the authors to adapt to meet changes from outside, always providing who wrote up the report you are holding in your hands or the profession’s contribution with constructive criticism of reading on screen and the Steering Group of Royal College what goes on within and increasingly beyond our hospitals. and Specialty Association nominees and others who are the Because it depends primarily on the old-fashioned altruism members of NCEPOD and who provided our peer review, and professionalism of clinicians who want to find out how criticising the raw data when it was presented to them and to do better for their patients, it is the one aspect of the NHS the report when early drafts were circulated. whose cost has not spiralled out of control. As medicine has become more complicated to deliver at the same time as As it is my last report I may also be allowed to mention our tolerance of variability has diminished, the need for that all my fellow Trustees that I have worked with over the professionalism has increased. In most cases we cannot make years: they have all been a pleasure to work with and the medicine, the most altruistic and unforgiving of professions, leadership they provide to the debates at the Steering Group better by regulation or by bringing in the criminal law. The has been especially stimulating. We have been a strongly mainstay of treatment must be to set the professionals free united group, all believing without any reservation in the and to give them the information that will enable them to work and effectiveness of the organisation. This is partly do better by their patients: that is all any of them want to because we have no doubt that we got the most important do, and the existence of NCEPOD – a nationwide organised single decision in my 11 years right, when we chose Dr process of self criticism that is composed of representatives Marisa Mason to be Chief Executive: she has been a superb of the professions - is evidence of their determination to leader and with Dr Neil Smith as her Deputy, and the improve what happens to all of us when we are ill. excellent NCEPOD staff behind them, this is a team that will go from strength to strength. It remains only for me to thank all those who have contributed to this study. The UK Sepsis Trust who asked us to look at the subject, the study advisory group who devised the detail of the study and told us what questions to ask; the Local Reporters and study contacts who spotted the cases and collected the notes and the clinicians who filled in the questionnaires; back at base the peer reviewers who Bertie Leigh did the heavy lifting of scrutinising the case notes and the NCEPOD Chair

8 Principal recommendations Back to contents

All hospitals should have a formal protocol for the early In line with previous NCEPOD and other national reports’ identification and immediate management of patients with recommendations on recognising and caring for the acutely sepsis. The protocol should be easily available to all clinical deteriorating patients, hospitals should ensure that their staff, who should receive training in its use. Compliance staffing and resources enable: with the protocol should be regularly audited. This protocol a. All acutely ill patients to be reviewed by a consultant should be updated in line with changes to national and within the recommended national timeframes (max of international guidelines and local antimicrobial policies. 14 hours after admission) (Medical Directors) b. Formal arrangements for handover c. Access to critical care facilities if escalation is required; An early warning score, such as the National Early Warning and Score (NEWS) should be used in both primary care and d. Hospitals with critical care facilities to provide a Critical secondary care for patients where sepsis is suspected. This Care Outreach service (or equivalent) 24/7. (Medical will aid the recognition of the severity of sepsis and can be Directors, Nursing Directors, Commissioners) used to prioritise urgency of care. (General Practitioners, Ambulance Trusts, Health Boards, NHSE, Clinical Directors, All patients diagnosed with sepsis should benefit from Royal Colleges) management on a care bundle as part of their care pathway. The implementation of this bundle should be audited and On arrival in the emergency department a full set of vital reported on regularly. Trusts/Health Boards should aim to signs, as stated in the Royal College of Emergency Medicine reach 100% compliance and this should be encouraged by standards for sepsis and septic shock should be undertaken. local and national commissioning arrangements. (Medical (Emergency Medicine Physicians, Clinical Directors, Nursing Directors, Clinical Directors, Commissioners) Directors) See the full list of recommendations on page 107

9 complications and adverse events

10 Introduction Back to contents

Sepsis is defined as an overwhelming response to infection specialties. When a patient has worsening vital signs they in which the immune system initiates a potentially need to be recognised and acted upon and whilst early damaging systemic inflammatory response syndrome (SIRS) warning scores such as NEWS are increasingly used9, the which can manifest in a number of physiological changes, possibility of sepsis should form part of that process. In recognised by worsening vital signs or ‘SIRS criteria’ 2010, the Scottish Trauma Audit Group (STAG) conducted (temperature, respiratory rate, heart rate). Severe sepsis is an audit of sepsis within acute hospital settings. 1.7% of defined as sepsis leading to dysfunction of one or more new admissions developed criteria for sepsis within 2 days organ systems according to current criteria.1 This year, of attendance; 34% of these patients met the criteria for international consensus definitions will be amended to focus severe sepsis, with a mortality of 24% in this group.10 on physiological changes of organ dysfunction, including hypotension, tachypnoea and altered mental state.2 Sepsis is Treatment of the infection in patients with sepsis is already recognised as difficult to diagnose and it can only be paramount. In 2010, the International Surviving Sepsis hoped that a new definition will aid this process. However, Campaign (SSC) published results in over 15,000 episodes whichever definition is used it is the wider consideration showing that delivery of early antibiotics (at that stage given to sepsis by healthcare professionals that is important. within 3 hours) was independently associated with survival, but was achieved in only 67% of cases.11 The Over 70% of cases of sepsis are believed to arise in the recommendation has since been changed to delivery of community.3 General practitioners and other pre-hospital antibiotics within 1 hour of severe sepsis being identified.1 services present key opportunities for prompt recognition However, the importance of administering antimicrobials in and treatment of sepsis. Patients requiring hospital care may an era when doctors are being advised not to over-prescribe be admitted through emergency departments or admissions them is somewhat confusing and this is an area that needs units, where the same issue of prompt recognition is attention to ensure that patients are treated effectively but equally important. In 2011, the Royal College of Emergency that there is robust antimicrobial stewardship.12,13 Medicine conducted an audit of compliance with sepsis management standards in emergency departments. One systematic issue that hinders the knowledge about Compliance was found to be suboptimal at 27-47%.4 A sepsis is its limited coding. Within the United Kingdom repeat audit in 2013-14 showed mixed results with marginal there is believed to be an underestimate of the incidence of improvement.5 sepsis as coding guidelines prioritise the source of infection over sepsis as a primary coded term. The incidence of Sepsis can also occur in patients already in hospital who severe sepsis depends on how acute organ dysfunction is acquire infections and whose condition deteriorates. In defined and on whether that dysfunction is attributed to 2005 NCEPOD reported that acutely ill patients were an underlying infection. Organ dysfunction is often defined languishing in wards not being recognised nor escalated by the need and provision of supportive therapy (e.g. quickly enough.6 Since then there have been National mechanical ventilation), and epidemiologic studies thus Institute for Health Excellence and Care (NICE) guidelines only count the cases in which treatment is undertaken. This produced (CG50)7 and work undertaken by the National under reporting of sepsis will mean that as a condition it Patient Safety Agency (NPSA as it was) around recognition will be under resourced, and there will be limitations in the of the critically ill patient.8 Sepsis is part of that severely opportunity to audit it and learn from the cases at mortality ill/deteriorating patient scenario and it is relevant to all reviews. In the UK an estimated 37,000 patients die with

11 Introduction

sepsis per year14 and a further estimate of 65,000 people Ombudsman published her first clinical report “Time to Act” per year survive episodes of severe sepsis, often with serious identifying common themes in 10 case studies of patients long-term complications: amputation, muscular contraction, who died following sepsis.21 This report identified failings irreversible damage to lungs, heart and kidneys, neuro- throughout the patient pathway: from carrying out a timely psychiatric disorders such as cognitive dysfunction and post- initial assessment and identifying the source of infection to traumatic stress disorder. Early recognition is therefore vital. adequate monitoring and timely initiation of treatment. This NCEPOD study similarly looks in detail at individual cases to There is an increasing focus on sepsis from health and identify common themes. 2013 also saw the formation of political organisations with a will to improve the care of the All Party Parliamentary Group (APPG) on sepsis which patients with sepsis. NHS England has identified tackling has recently published 10 recommendations in a report sepsis as a clinical priority for improving patient outcomes highlighting similar themes to those presented here.22 for 2015/16.15 Sepsis has been linked to a new CQUIN in England.16 NICE are currently developing sepsis guidelines.17 Sepsis is a major cause of avoidable mortality and morbidity. A new study is assessing the ‘Size of Sepsis in Wales’18 This study, whilst considering the plethora of other work following on from a point-prevalence study in 2014.19 In in this important area, sets out to identify in greater detail, 2014 MBRRACE-UK published a themed confidential enquiry remediable factors which if addressed would improve the which reviewed maternal mortality and morbidity due quality of care of patients with sepsis. to sepsis.20 In 2013 the Parliamentary and Health service

12 1

Method and Data Returns Back to contents

Method Hospital participation National Health Service hospitals in England, Wales and Study Advisory Group Northern Ireland were expected to participate as well as The Study Advisory Group comprised a multidisciplinary hospitals in the independent sector and public hospitals group of senior clinicians from the following specialties: in the Isle of Man, Guernsey and Jersey. Within each acute medicine, emergency medicine, general surgery, hospital, a named contact, referred to as the NCEPOD Local obstetrics and gynaecology, microbiology, critical care Reporter, acted as a link between NCEPOD and the hospital medicine, pathology, public health strategy, general practice, staff, facilitating case identification, dissemination of critical care outreach nursing and patient representation. questionnaires and data collation.

Study aim Population The aim of the study was to identify and explore remediable Adult patients, ≥16 years old, identified as being seen factors in the process of care for patients with sepsis. by the Critical Care Outreach Team or equivalent, or who were admitted directly to critical care during the study Objectives period with a diagnosis of sepsis, based on presence of • To examine organisational structures, processes, infection, documented or suspected, and two or more of protocols and care pathways for sepsis recognition and the following: management in hospitals from admission through to • Fever (> 38.3°C)/hypothermia (core temperature discharge or death. < 36°C) • To identify avoidable and remediable factors in the • Heart rate > 90/min – 1 or more than two standard management of the care for a representative sample deviations above the normal value for age of adult patients with sepsis, throughout the patient • Tachypnoea (respiratory rate >20 breaths/minute) pathway from presentation to primary care (if • Acutely altered mental status applicable), throughout secondary care to discharge or • Arterial hypotension (systolic blood pressure < 90 death, focusing on the following areas of care: mmHg, mean arterial pressure < 70 mmHg, or a systolic * Evaluation of the use of systems and processes blood pressure decrease > 40 mmHg or less than two that are in place within hospitals to facilitate timely standard deviations below normal for age) identification, escalation and appropriate treatment • Hyperglycemia (plasma glucose > 140 mg/dL or 7.7 of infection, including transfer to high dependency mmol/L) in the absence of diabetes and intensive care units where appropriate • Leukocytosis (white blood cell count > 12,000 μL–1) or * Examining the recognition of sepsis and early signs Leukopenia (white blood cell count < 4000 μL–1) (or of septic shock across the entire patient pathway normal white blood cell count with >10% immature * Investigating the appropriate management of sepsis forms) * Reviewing whether there was a multidisciplinary team approach Adapted from: Signs & symptoms of infection highlighted in * Assessing the adequacy of communication with Surviving Sepsis Campaign Sepsis Screening Tool.1 families and carers, as could be ascertained from the case notes From the cases identified, a sample of 5 cases per hospital * Examining the management of the end of life was randomly selected to be included in the study. pathway and ceilings of treatment

13 Method and Data Returns

Exclusions Questionnaires and case notes • Immunosuppressed neutropaenic patients on chemotherapy, immunosuppressant drugs or patients Two questionnaires were disseminated to collect data for with solid organ transplant this study; a clinician questionnaire relating to each patient • Pregnant women up to 6 weeks post-partum (covered included and an organisational questionnaire for each by MBRRACE-UK maternal sepsis morbidity study)20 hospital participating in the study, regardless of whether • Patients on an end of life care pathway at the time of they had patients included in the study. Questionnaires diagnosis, or a consultant-led decision made not to were designed with input from the Study Advisory Group. escalate (prior to entry into the study) • Patients who developed sepsis after 48 hours on Clinician questionnaire critical care This questionnaire was sent to the named consultant responsible for the patient prior to admission to critical Case identification care/ review by the Critical Care Outreach Team. If the During the two-week data collection period, 6th-20th consultant was not the most suitable person to complete May 2014, all patients who met the inclusion criteria were the questionnaire then they were asked to identify a more identified prospectively by nominated study contacts in appropriate consultant. This questionnaire was used to critical care and on the Critical Care Outreach Team. collect data on the care of the patient throughout their pathway of care from presentation with sepsis to death, Whilst it was assumed that prospectively identifying discharge or remaining in hospital 30 days after admission. patients with sepsis is more effective than relying on a retrospective identification through ICD10 coding, it is Organisational questionnaire possible that the Study Contacts on critical care and the The organisational questionnaire was sent to the NCEPOD Critical Care Outreach Team may not have identified every Local Reporter to be completed with the help of relevant possible patient that was eligible for the study. However, specialty leads. Data were requested on the policies and this would not have affected the sampling as the peer protocols in place at each hospital, on the availability of reviewed sample was limited to a maximum of five cases services, facilities and staffing relevant to patients with per hospital anyway. This study is a snapshot of the care sepsis. Information was also collected on any sepsis care provided to patients with sepsis. quality improvement initiatives.

Furthermore, this study was designed to examine the care In addition to the acute hospitals to which patients with of patients who were more unwell with sepsis, by only sepsis would be admitted for treatment, community including patients who were either admitted to critical care and independent hospitals were also included in the or who were reviewed by the Critical Care Outreach Team. organisational part of the study, despite the fact that they The study was not therefore able to comment on the care may not have patients with sepsis admitted to them. This of patients who died in the community or in the emergency was to see if there were organisational structures in place department, or who died before being reviewed by the to manage the initial care of patients who may develop Critical Care Outreach Team. Nor can comment be made sepsis as an inpatient. on those patients who were never escalated to the Critical Care Outreach Team or critical care, either because they had timely interventions and did not deteriorate sufficiently or because they had treatment limitation decisions made early in their pathway or those patients who were never recognised as having sepsis.

14 1

Case notes Reviewers answered a number of specific questions by Photocopied case note extracts were requested for each direct entry into a database, and were also encouraged to case that was to be peer reviewed. For the entire admission: enter free text commentary at various points. • All inpatient annotations/medical notes • Nursing notes The grading system below was used by the Reviewers to • Critical care notes grade the overall care each patient received: • Operation/procedure notes • Anaesthetic charts Good practice: A standard that you would accept from • Observation charts yourself, your trainees and your institution. • Haematology/biochemistry/microbiology results Room for improvement: Aspects of clinical care that • Fluid balance charts could have been better. • Drug charts Room for improvement: Aspects of organisational • Consent forms care that could have been better. • Discharge letter/summary Room for improvement: Aspects of both clinical and • Autopsy report if applicable organisational care that could have been better. Less than satisfactory: Several aspects of clinical and/or General practitioner (GP) case notes organisational care that were well below that you would For cases where it was recorded on the clinician accept from yourself, your trainees and your institution. questionnaire that the patient had been seen by their GP in Insufficient data: Insufficient information submitted to relation to the hospital admission for sepsis (regardless of NCEPOD to assess the quality of care. whether or not the GP referred the patient to hospital), the details of the GP were extracted from the case notes and copies of the GP case notes for the two-week period prior Quality and confidentiality to the hospital admission were requested. Each case was given a unique NCEPOD number. The data Peer review from all questionnaires received were electronically scanned into a preset database. Prior to any analysis taking place, A multidisciplinary group of Reviewers was recruited the dataset was cleaned to ensure that there were no to peer review the case notes and associated clinician duplicate records and that erroneous data had not been questionnaires. The group of Reviewers comprised entered during scanning. Any fields containing data that consultants, associate specialists, trainees and clinical nurse could not be validated were removed. Section 251 approval specialists, from the following specialties: acute medicine, had been granted for this study. emergency medicine, general medicine, nephrology, critical care outreach, anaesthesia, intensive care medicine, Data analysis respiratory medicine, microbiology and general and plastic surgery. In addition general practitioners were recruited to Following cleaning of the quantitative data, descriptive review the GP case notes separately. data summaries have been produced. The qualitative data collected from the Reviewers’ opinions and free text Questionnaires and case notes were anonymised by the answers in the clinician questionnaires were coded, where non-clinical staff at NCEPOD prior to peer review. After applicable, according to content to allow quantitative being anonymised each case was reviewed by at least analysis. The data were reviewed by NCEPOD Clinical one Reviewer within a multidisciplinary group. At regular Co-ordinators, a Clinical Researcher and a Researcher, to intervals throughout the meeting, the Chair allowed a identify the nature and frequency of recurring themes. period of discussion for each Reviewer to summarise their cases and ask for opinions from other specialties or raise aspects of the case for discussion.

15 Method and Data Returns

Case studies have been used throughout this report to to assess 551 cases, the remainder of the returned case illustrate particular themes. note extracts were either too incomplete for assessment or were returned after the final deadline and last case reviewer All data were analysed using Microsoft AccessTM and ExcelTM meeting. There were 129 cases identified where the patient by the research staff at NCEPOD. saw their GP in relation to the admission (which did not necessarily lead to a referral to hospital). Of these, 60 sets of The findings of this report were reviewed by the Study GP notes were received and 54 were suitable for review. Advisory Group, Reviewers, NCEPOD Clinical Co-ordinators and the NCEPOD Steering Group prior to publication. Study sample denominator by chapter

Data returns Within this study the denominator will change for each chapter and occasionally within each chapter. This is In total 3,363 patients from 305 hospitals were identified because data have been taken from different sources as meeting the study inclusion criteria during the two-week depending on the analysis required. For example, in some case identification period (Figure 1). When the sampling cases the data presented will be a total from a question criterion of 5 cases per hospital was applied, 884 cases taken from the clinician questionnaire only, whereas some were selected for inclusion. A total of 710/884 (80.3%) analysis may have required the clinician questionnaire and completed clinician questionnaires and 657 sets of case the Reviewers’ view taken from the case notes. notes were returned to NCEPOD. The Reviewers were able

Number of cases indentified within the two-week study period n=3363

Number of cases selected for inclusion n=884

Number of questionnaires Number of sets of case notes returned n=710 returned n=657

Number of questionnaires Number of cases peer included in the reviewed analysis n=710 n=551

Number of cases indentified with GP input and GP details - request sent for notes n=129

Number of GP case notes returned n=54

Figure 1 Data returns

16 2

Organisational data Back to contents

An organisational questionnaire was sent to all participating Table 2.1 Hospital type hospitals to understand the systems and processes in Type Number of % place to manage patients with sepsis. In all, 549 hospitals hospitals responded (Table 2.1) with 162 (29.5%) hospitals identified District General Hospital (DGH) 102 18.6 as District General Hospitals (DGH) and 53 (10%) as ≤ 500 beds University Teaching Hospitals (UTH). There was also a District General Hospital (DGH) 60 10.9 good response from those identified as independent and > 500 beds community hospitals (however, it can be seen later that University Teaching Hospital (UTH) 53 9.7 these hospitals contributed few patients to this study since Tertiary Specialist Centre (TSC) – 28 5.1 only hospitals with on-site critical care facilities participated stand alone in the case review part of the study). The remaining hospitals Independent Hospital (IH) 83 15.1 from which a response was received were described as a non- Community or Cottage Hospital (CH) 204 37.2 acute peripheral or satellite hospital within an acute Trust/ Peripheral Hospital (PH) 6 1.1 Health Board or a rehabilitation hospital. Rehabilitation Hospital (RH) 13 2.4

Of the hospitals from which a response was received, 201 Total 549 hospitals reported that they had an emergency department (ED). They comprised 149 DGHs and 40 UTHs (Table 2.2).

Table 2.2 Availability of an emergency department Available Yes No Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 90 11 101 1 102 District General Hospital (DGH) > 500 beds 59 1 60 0 60 University Teaching Hospital (UTH) 40 13 53 0 53 Tertiary Specialist Centre (TSC) – stand alone 5 23 28 0 28 Independent Hospital (IH) 1 77 78 5 83 Community or Cottage Hospital (CH) 5 171 176 28 204 Peripheral Hospital (PH) 1 5 6 0 6 Rehabilitation Hospital (RH) 0 11 11 2 13 Total 201 312 513 36 549 (39.2%) (60.8%)

17 organisational data

Table 2.3 Opening hours of the emergency Protocols for identification and management department of sepsis Opening hours Number of % hospitals Early recognition and management of sepsis leads to 23,24 24 hours/day, 7 days/week 192 96.0 reduction in morbidity and mortality and administration of Normal working hours (8am-6pm) 1 <1 an effective antimicrobial within the first hour of hypotension 7 days/week has been associated with a survival rate of 79.9%.25,26 Other hours 7 3.5 Care bundles and protocols have made a major impact Subtotal 200 in management of time sensitive medical conditions like Not answered 1 stroke and heart attacks.27 Such protocols and bundles have Total 201 also been shown to help with the early identification and management of sepsis.28 They will be discussed later in the A large majority of hospitals reported they had an ED which report at the patient level, but at an organisational level provided a 24/7 service (Table 2.3). The seven hospitals that 360/544 (66%) hospitals reported that they had a sepsis did not provide such services were all community hospitals protocol. Within a subgroup of DGHs and UTHs 194/212 that provided extended twilight cover usually up to 8pm (92%) had a protocol, while 50% of the other hospitals had or 10pm. one (Table 2.4).

Table 2.4 Presence of a specific protocol/care pathway/ bundle for recognition and management of patients with sepsis Sepsis protocol Yes No Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 91 10 101 1 102 District General Hospital (DGH) > 500 beds 55 5 60 0 60 University Teaching Hospital (UTH) 48 3 51 2 53 Tertiary Specialist Centre (TSC) – stand alone 18 9 27 1 28 Independent Hospital (IH) 54 29 83 0 83 Community or Cottage Hospital (CH) 79 124 203 1 204 Peripheral Hospital (PH) 4 2 6 0 6 Rehabilitation Hospital (RH) 11 2 13 0 13 Total 360 184 544 5 549 (66.1%) (33.8%)

Table 2.5 Source of protocol There was some uniformity in what was included in hospital protocols and 4 out of 5 hospitals drew on national Source of protocol Number of % hospitals and international guidance to produce their local protocols (Table 2.5). Taken directly from national/ 92 26.8 international guidelines Modified version of national/ 208 60.6 international guidelines Locally developed protocol/care 66 19.2 pathway/ bundle

Answers may be multiple n=343; not answered in 17

18 2

Table 2.6 Details of protocol Actions included: Yes % No % Subtotal Not Total answered Administering IV fluids 330 94.3 19 5.4 349 11 360 Administering IV antimicrobials 323 95.3 16 4.7 339 21 360 Blood cultures to be taken before 320 93.6 22 6.4 342 18 360 antimicrobials administered Administering oxygen therapy 328 95.1 17 4.9 345 15 360 Early lactate measurement 307 91.4 29 8.6 336 24 360 Catheterisation/urine output measurement 323 93.9 21 6.1 344 16 360 Answers may be multiple

Table 2.6 shows some of the details found in the Most protocols did specify a timeframe for actions to be protocol, the 16 hospitals whose protocol did not include taken on identifying sepsis. All hospitals responded that administering IV antimicrobials were all community the actions would be completed within six hours and the hospitals. majority actually set this to less than one hour (Table 2.8).1

The questionnaire collated data on whether a time frame Of note is that 11 of the 184 hospitals that reported was specified for completing these early actions detailed in that they did not have a protocol for identifying and the protocols (Table 2.7).1 managing sepsis did not have a more general protocol for management of the deteriorating patient. These included Table 2.7 Specified time frame for actions listed in one DGH, two TSCs, one IH and seven CHs (Table 2.9). sepsis protocol Time frame Number of % Table 2.9 Availability of a general protocol for the hospitals management of deteriorating patients in hospitals with no sepsis protocol Yes 325 94.2 No 20 5.8 General protocol Number of % hospitals Subtotal 345 Yes 154 93.3 Not answered 15 No 11 6.7 Total 360 Subtotal 165 Table 2.8 Time frame for actions listed in sepsis Not answered 19 protocol Total 184 Time frame Number of % hospitals Since easy access to hospital protocols is important, 0 - 1 hours 305 95.0 respondents were asked about the availability/ accessibility > 1 - 6 hours 16 5.0 of protocols, policies and guidelines. In 97% (518/532) of hospitals the protocols, policies and guidelines were Subtotal 321 available on the hospital intranet and therefore easily Not answered 4 accessible during all hours (Table 2.10 overleaf). However, Total 325 despite the reported easy availability, when reviewing how the antimicrobial therapy was chosen (discussed in detail in Chapter 7); only 36% of patients received the drug specified by the local hospital policy.

19 organisational data

Table 2.10 Access to protocols If hospitals have a protocol to aid the recognition and early management of sepsis, the education of staff members in Availability Number of % hospitals its use is vital. The majority of hospitals reported that some form of sepsis training for new members of medical and Printed copies stored in relevant 253 47.6 locations nursing staff was provided, but less commonly outside of Electronic copies on hospital 518 97.4 the emergency department. (Table 2.11) intranet Internet 91 17.1 Other 8 1.5 Answers may be multiple n=532; not answered in 6

Table 2.11 Provision of education around sepsis recognition and management, including the use of the protocol for hospital staff Education regarding sepsis Medical staff Nursing staff Emergency % Other % Emergency % Other % department wards department wards Yes 149 84.2 240 78.7 150 83.3 228 72.6 No 28 15.8 65 21.3 30 16.7 86 27.4 Subtotal 177 305 180 314 Not answered 32 55 143 46 Not applicable 151 0 37 0 Total 360 360 360 360

Table 2.12 Early warning score linked to escalation Escalation of care protocols Early warning score linked to Number of % The vast majority of respondents (530/538; 99%) reported escalation hospitals that some form of track and trigger tool was employed to Yes 516 97.9 monitor deteriorating patients (data not shown), and most No 11 2.1 stated that their early warning scoring systems were linked to their escalation protocols (Table 2.12). This is consistent Subtotal 527 with previous NCEPOD findings.29 Not answered 3

Total 530 Table 2.13 details the actions triggered by the escalation protocol. In 233/515 (45%) hospitals there appeared to be a system in place to provide critical care expertise to the treating team in the event of the escalation protocol being triggered, either from the Critical Care Outreach Team or through a direct referral to critical care or both (data not shown).

20 2

Table 2.13 Detail of escalation protocol Escalation (n/100) (n/57) (n/49) (n/26) (n/80) (n/184) (n/6) (n/13) Total protocols DGH DGH UTH TSC IH CH PH RH involve: ≤ 500 > 500 beds beds Call to critical care 73 48 32 19 16 4 4 0 196 outreach team 38.1% Call to rapid 5 5 8 3 3 0 0 0 24 response team 4.7% Level 3 referral 58 29 29 12 19 7 3 0 157 30.5% Call to medical 29 21 16 7 14 11 3 3 104 emergency team 20.2% Call to cardiac 30 19 18 9 26 11 3 1 117 arrest team 22.7% Transfer to other 8 2 6 8 39 157 3 6 229 hospital 44.5% Review by patient's 91 49 46 21 69 117 5 11 410 own medical team 79.6% Other 11 7 7 4 19 60 1 5 115 22.3% Answers may be multiple n=515; not answered=1

Table 2.14 Pre-alert system for incoming sepsis patients Pre-alert system Yes No Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 44 39 83 10 93 District General Hospital (DGH) > 500 beds 25 30 55 5 60 University Teaching Hospital (UTH) 24 17 41 8 49 Tertiary Specialist Centre (TSC) – stand alone 2 5 7 17 24 Total 95 91 186 40 226 (51.1%) (48.9%)

Respondents were also asked if there was a system where a pre-hospital alert could be sent by a GP or ambulance crew, warning of the arrival of a patient with suspected sepsis in order that their management could be expedited. Of the 186 hospitals from which a response was received, 95 (51%) reported that such a system existed (Table 2.14). These data are displayed for the 226 acute hospitals to which patients could have been admitted as an emergency.

21 organisational data

Table 2.15 Microbiology input (Acute hospitals) Clinical microbiology ward rounds - Level 3 Level 2 General General Other locations and frequency medical surgical inpatient ward ward ward Daily 139 115 20 13 18 % 68.1 57.2 10.3 6.6 9.0 Bi-weekly 26 26 17 17 19 % 12.7 12.9 8.8 8.7 9.5 Weekly 12 16 35 32 44 % 5.9 8.0 18.0 16.3 22.1 Telephone support only 3 16 50 56 50 % 1.5 8.0 25.8 28.6 25.1 Other 23 25 49 49 46 % 11.3 12.4 25.3 25.0 23.1 No microbiology ward round 1 3 23 29 22 % 0.5 1.5 11.9 14.8 11.1 Subtotal 204 201 194 196 199 N/A 15 11 16 13 - Not answered 7 14 16 17 27 Total 226 226 226 226 226

Administration of antimicrobials The 23 hospitals falling under the group ‘other’ included seven that had a weekday service only, 12 provided ward rounds Early identification of infection and administration of three times per week, and two had four ward rounds each appropriate antimicrobials is vital in managing sepsis. week. Two other hospitals had provision for microbiology input Regular review and de-escalation of treatment (antimicrobial as required but did not have formal ward rounds. stewardship) also plays an important role in preventing complications and improving recovery. Hospitals were The vast majority of acute hospitals (224/226) had a policy asked about the availability of microbiology input available for antimicrobial use (data not shown). Respondents (Table 2.15). were also asked if there was a policy on who was able to administer intravenous antimicrobials (Table 2.16).

Table 2.16 Policy for staff that can administer antimicrobials Policy Yes No Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 69 20 89 4 93 District General Hospital (DGH) > 500 beds 47 12 59 1 60 University Teaching Hospital (UTH) 33 12 45 4 49 Tertiary Specialist Centre (TSC) – stand alone 16 7 23 1 24 Total 165 51 216 10 226 (76.4%) (23.6%)

22 2

Of the 216 acute hospitals from which a response was Transfer to critical care received, 165 (76.4%) had a policy for the administration of IV antimicrobials (Table 2.16). Multiple staff members were Of the 272 hospitals that did not have critical care facilities able to give antimicrobials, with staff nurses being the most on site 22% did not have formal arrangements to transfer common group (Table 2.17). critically ill patients, almost all of them were community hospitals (Table 2.18). Table 2.17 Detail of staff who can administer intravenous antimicrobials (Acute hospitals only) For those hospitals where the policy was to transfer patients Staff Number of % with sepsis to another hospital for management, we aimed hospitals to assess the duration of an average journey in the middle Senior doctor (ST3 or above) 133 81.1 of the day. Only in five hospitals was the time greater than Junior doctor (below ST3) 135 82.3 one hour. The majority reported that the time was less than 30 minutes (Figure 2.1 overleaf). Other healthcare worker 37 22.6 Senior nurse (senior staff nurse or 140 85.4 above) In cases where the patients had to be transferred to an alternative site for critical care, the clinical interventions Staff nurse 125 76.2 undertaken prior to transfer were assessed. The ‘Surviving Healthcare assistant 3 1.8 Sepsis Campaign’ and others have recommended the use Answers may be multiple n=164; not answered in 1 of the early resuscitation bundle. The steps taken prior to transfer are summarised in Table 2.19 overleaf. These were routinely undertaken in only half of the hospitals.

Table 2.18 Hospitals without critical care on-site – critical care transfer arrangement exists with nearby hospital(s) Critical care transfer arrangement Yes No Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 11 0 11 0 11 University Teaching Hospital (UTH) 4 0 4 0 4 Tertiary specialist centre (TSC) – stand alone 5 1 6 0 6 Independent Hospital (IH) 33 1 34 0 34 Community or Cottage Hospital (CH) 130 55 185 14 199 Peripheral Hospital (PH) 5 0 5 0 5 Rehabilitation Hospital (RH) 13 0 13 0 13 Total 201 57 258 14 272 (77.9%) (22.1%)

23 organisational data

Pecentage of hospitals

50 81 45

40

35

30 47 25

20

15

18 10 11 12 8 5 5

0 Under 10 mins 10-20 mins 20-30 mins 30-40 mins 40-50 mins 50-60 mins Over 60 mins

Figure 2.1 Time to transfer to critical care if critical care was not on-site (n=182, missing data in 90)

Table 2.19 Steps performed prior to transfer for off-site critical care Steps taken (n/11) (n/4) (n/5) (n/33) (n/130) (n/5) (n/13) Total DGH UTH TSC IH CH PH RH ≤ 500 beds Take blood cultures 5 3 4 27 52 4 4 99 56.9% Administer 6 3 4 20 42 4 5 84 48.3% antimicrobials Administer oxygen 7 3 5 28 84 4 5 136 78.2% therapy Haemodynamically 7 3 4 26 34 4 3 81 46.6% stabilise the patient (fluids) Measure lactate 4 3 3 17 7 3 1 38 21.8% Attempt to isolate 1 2 3 9 31 1 3 50 28.7% the source of infection Monitor urine 6 3 4 27 57 4 4 105 60.3% output Other 0 0 0 4 19 0 1 24 13.8% None 1 0 0 2 25 0 2 30 17.2% Answers may be multiple n=174; not answered in 27

24 2

Table 2.20 Specific proforma to monitor progress of patients with sepsis (Acute hospitals only) Sepsis proforma Yes No Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 28 61 89 4 93 District General Hospital (DGH) > 500 beds 14 45 59 1 60 University Teaching Hospital (UTH) 10 34 44 5 49 Tertiary Specialist Centre (TSC) – stand alone 3 20 23 1 24 Total 55 160 215 11 226 (25.6%) (74.4%)

Table 2.21 Hospital had a care bundle for source isolation/control Care bundle for source Yes No Subtotal NA - source Not Total control control not answered carried out District General Hospital (DGH) 16 71 87 9 6 102 ≤ 500 beds District General Hospital (DGH) 8 44 52 1 7 60 > 500 beds University Teaching Hospital 8 37 45 0 8 53 (UTH) Tertiary specialist centre (TSC) – 5 17 22 4 2 28 stand alone Independent Hospital (IH) 12 28 40 30 13 83 Community or Cottage Hospital 31 68 99 84 21 204 (CH) Peripheral Hospital (PH) 1 3 4 1 1 6 Rehabilitation Hospital (RH) 1 7 8 4 1 13 Total 82 275 357 133 59 549 (23%) (77%)

Since there is evidence that using checklists, protocols Identifying the source of infection and proformas improve the safety and quality of care, respondents were asked if they used a proforma to monitor Identifying and treating the source of infection is one of progress once sepsis was diagnosed (Table 2.20). Only 55 of the definitive measures in the medical management of the acute hospitals that responded (26%) reported that such sepsis after initial resuscitation. Only 82/357 hospitals a proforma was being used. (23%) confirmed that a care bundle/protocol existed for this purpose. Of the DGHs and UTHs, 32/184 (17%) had such a protocol (Table 2.21).

25 organisational data

Handovers

Good communication, safe and efficient handovers are important factors in improving quality of care. Over 40% of hospitals reported that they did not have a policy on handovers (Table 2.22).

Table 2.22 Policy for staff handover (All hospitals) Policy Yes No Subtotal Not Total answered DGH ≤ 500 beds 70 27 97 5 102 DGH > 500 beds 39 17 56 4 60 University teaching hospital 33 13 46 7 53 Tertiary specialist centre - stand alone 17 9 26 2 28 Independent hospital 36 43 79 4 83 Community or cottage hospital 93 95 188 16 204 Peripheral Hospital (PH) 6 0 6 0 6 Rehabilitation Hospital (RH) 6 7 13 0 13 Total 300 211 511 38 549 (58.7%) (41.3%)

Table 2.23 Details of the staff handover policy Policy includes Clear escalation Structured Time set aside plan proforma for for face to face handover handover Number of % Number of % Number of % hospitals hospitals hospitals Yes 212 74.9 230 78.5 270 94.1 No 71 25.1 63 21.5 17 5.9 Subtotal 283 293 287 Not answered 17 7 13 Total 300 300 300

Table 2.24 Patients provided with printed Of the acute hospitals, 63/226 (28%) did not have a policy information about sepsis (All hospitals) (data not shown). However, for those hospitals that had a Patients information about Number of % policy, there was a good level of detail (Table 2.23). sepsis hospitals Yes 29 5.6 On enquiring about discharge process, especially No 490 94.4 information provided to patients, it was of note that fewer than 6% (29/519) of hospitals had a robust system Subtotal 519 to ensure that useful information relevant to sepsis was Not answered 30 provided (Table 2.24). Total 549

26 2

Table 2.25 Follow-up service for patients post discharge (Acute hospitals only) Follow-up service Yes No Subtotal Not Total answered DGH ≤ 500 beds 32 59 91 2 93 DGH > 500 beds 21 39 60 0 60 University Teaching Hospital 20 21 41 8 49 Tertiary Specialist Centre - stand alone 5 18 23 1 24 Total 78 137 215 11 226 (36.3%) (63.7%)

Follow-up study that Critical Care Outreach Teams were available in 199 of 223 (89%) hospitals with critical care from which a Sepsis is recognised to lead to long-term physical and response was received (Table 2.26). neuropsychological conditions.30 A follow-up service would therefore be of benefit to patients and has been Table 2.26 Critical Care Outreach Team or equivalent (Hospitals with critical care) recommended by NICE Clinical Guideline 83 for the follow- up of critical care discharges.31 Only 78/215 (36.3%) Critical care outreach team or Number of % respondents confirmed that they provided a follow-up equivalent hospitals service (Table 2.25). Yes 199 89.2 No 24 10.8 Critical Care Outreach Teams Subtotal 223 Not answered 1 The NCEPOD report “An Acute Problem?” was published in Total 224 20056 and reported on the care of medical patients referred to critical care units. The report showed that only 92 of 208 Hospitals in which a Critical Care Outreach Team (24/223) (44.2%) hospitals studied had Critical Care Outreach Teams was not employed, included both small and large hospitals and recommended that they should be available round the (Figure 2.2). clock each day in all hospitals. It is of note therefore in this

Number of hospitals Yes No

90

80

70

60

50

40

30

20

10

0 DGH ≤ 500 beds DGH > 500 beds UTH TSC

Figure 2.2 Hospital type and presence of a Critical Care Outreach Team

27 organisational data

Of the 199 hospitals that stated a Critical Care Outreach The system of coverage provided by Critical Care Outreach Team was present, the teams were staffed primarily with Teams varied across hospitals with 49% (96/196) providing a nurses from the critical care environment (Table 2.27). service 24/7 (Table 2.29).

Table 2.27 Grade of clinicians in the Critical Care Table 2.29 Availability of Critical Care Outreach Outreach Team (Acute hospitals) Teams (or equivalent) Grade Number of % Availability Number of % hospitals hospitals Consultant 56 28.4 24 hours, 7 days/week 96 49.0 Trainee 24 12.1 Normal working hours (8am-6pm) 26 13.3 Nurses 192 97.4 7 days/week Normal working hours (8am-6pm) 14 7.1 Answers may be multiple n=197; not answered in 2 Mon-Fri Over a quarter (28%) included a consultant, and a smaller Extended working hours, 45 23.0 7 days/week number (12%) had trainees (Table 2.28). Extended working hours, Mon-Fri 3 1.5 Table 2.28 Specialty of clinicians in the Critical Care Extended working hours, Mon-Fri 2 1.0 Outreach Teams (or equivalent) + reduced cover on weekends Speciality Number of % Other 10 5.1 hospitals Subtotal 196 Critical care 156 89.7 Not answered 3 Acute care 31 17.8 Total 199 Other 19 10.9 Answers may be multiple n=174; not answered in 25

Percentage

100

90

80

70

60

50

40

30

20

10

0 Emergency Medical Surgical General wards Specialist wards Other department assessment unit assessment unit

Figure 2.3 Areas covered by the Critical Care Outreach Team (n=197; not answered in 2)

28 2

The Critical Care Outreach Teams were found to provide Critical Care Outreach Teams were also reported to take their services to most clinical inpatient areas within the referrals through automated systems triggered by early hospital (Figure 2.3). ‘Other’ areas covered included the warning score thresholds (Table 2.30). For the ‘other’ group, cardiac catheterisation laboratory (4), and day surgery and the Critical Care Outreach Team took referrals from staff on recovery units (2). the ward, both medical and nursing.

Table 2.30 Trigger to call the Critical Care Outreach Consultant review Team or equivalent Call trigger Number of % There is increasing evidence that the presence of a hospitals consultant improves early identification and management Automated system linked to 28 14.1 of serious illness.32 Figure 2.4 summarises the consultant monitoring cover in different areas of the hospital. Some respondents Concern expressed by medical staff 189 95.5 stated that they had consultant presence on-site 24 hours a Concern expressed by nursing staff 187 94.4 day in the acute and critical care environment. Many other Early warning score 184 82.9 hospitals also provided extended cover by a consultant over Other 31 15.7 the evenings and weekends. Answers may be multiple n=198; not answered=1

Percentage of hospitals On site 24/7 Normal working hours (8am-6pm), Mon-Fri Normal working hours 100 (8am-6pm), 7 days/week Normal working hours plus on-call consultant 90 (weekend/nights)

80

70

60

50

40

30

20

10

0 Emergency department Medical assessment unit Surgical assessment unit Critical (ICU/HDU) n=163 n=178 n=161 n=199 Figure 2.4 Consultant cover in different areas of the hospital (All hospitals)

29 organisational data

Table 2.31 Lead clinician responsible for improving care of patients with sepsis (All hospitals) Lead clinician for sepsis Yes No Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 51 28 79 14 93 District General Hospital (DGH) > 500 beds 39 17 56 3 59 University Teaching Hospital (UTH) 28 16 44 1 45 Tertiary specialist centre (TSC) – stand alone 3 8 11 6 17 Independent Hospital (IH) 3 27 30 16 46 Community or Cottage Hospital (CH) 37 54 91 44 135 Peripheral Hospital (PH) 2 2 4 1 5 Rehabilitation Hospital (RH) 3 4 7 1 8 Total 166 156 322 86 408 (51.6%) (48.4%)

From the returned questionnaires it could be seen that Table 2.32 Designated sepsis nurse 408/549 hospitals had some form of quality improvement Sepsis nurse Number of % activity for sepsis. Clinical champions are beneficial in hospitals improving quality of care and over half the hospitals in Yes 38 10.7 the study (166/322; 52%) had appointed a Lead Clinician No 317 89.3 for sepsis (Table 2.31) and just over 10% (38/355) had a Subtotal 355 designated sepsis nurse (Table 2.32). Not answered 53 Sepsis response kit Total 408

The use of a sepsis response kit, bag or trolley can help deliver or trolley was being used (Table 2.33). The kits were located the sepsis early care bundle in the shortest time possible. A in the ED in 89% of hospitals, acute wards in 69% and total of 112 hospitals reported that a sepsis response kit, bag general wards in 79% of hospitals (data not shown).

Table 2.33 Sepsis response kit, bag or trolley (All hospitals) Sepsis trolley Yes No Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 35 55 90 3 93 District General Hospital (DGH) > 500 beds 21 38 59 0 59 University Teaching Hospital (UTH) 22 23 45 0 45 Tertiary Specialist Centre (TSC) – stand alone 5 10 15 2 17 Independent Hospital (IH) 15 31 46 0 46 Community or Cottage Hospital (CH) 11 121 132 3 135 Peripheral Hospital (PH) 1 4 5 0 5 Rehabilitation Hospital (RH) 2 6 8 0 8 Total 112 288 400 8 408 (28.0%) (72.0%)

30 2

Table 2.34 Number of serious incidents involving an episode of severe sepsis (All hospitals) Number of serious incidents 0 1-10 Subtotal Not Total answered District General Hospital (DGH) ≤ 500 beds 20 15 35 67 102 District General Hospital (DGH) > 500 beds 5 17 22 38 60 University Teaching Hospital (UTH) 14 14 28 25 53 Tertiary Specialist Centre (TSC) – stand alone 10 4 14 14 28 Independent Hospital (IH) 62 3 65 18 83 Community or Cottage Hospital (CH) 49 1 50 154 204 Peripheral Hospital (PH) 1 1 2 4 6 Rehabilitation Hospital (RH) 6 0 6 7 13 Total 167 55 222 327 549 (75.2%) (24.8%)

Serious incidents SIs was low. The low number is likely to be due to the poor documentation of sepsis or severe sepsis as a diagnosis The term Serious Incident (SI) is an incident that results in which would be improved with better coding. These data unexpected or avoidable death or severe harm. Investigating therefore, should not be over-interpreted (Table 2.34). such incidents can help in improving hospital systems and prevent recurrence of harm by ensuring that key lessons are Audit identified and learned. Information was collected on whether there was an audit in Respondents were asked how many SIs, where sepsis was acute hospitals of the number of episodes of sepsis where thought to have contributed, had been recorded between antimicrobials were received within one hour of diagnosis April 2013 to March 2014. This question was frequently not of sepsis (Table 2.35). Less than 50% of acute hospitals that answered (n=327), and overall the number of sepsis related responded carried out such audits.

Table 2.35 Audit of antimicrobial delivery (Acute hospitals) Hospital audits number Yes % No % Subtotal Not Total of episodes of sepsis answered where patient receives antimicrobials within the first hour of: Severe sepsis identification 90 44.1% 114 55.9% 204 22 226 Sepsis identification 75 36.9% 128 63.1% 203 23 226 Other identification 32 24.2% 100 75.8% 132 94 226 Answers may be multiple

31 organisational data

In addition, about 20% of respondents reported that they hospital audit or database (n=13), using the ICNARC had a system in place to record incidents of sepsis and reporting system (n=10) or monitoring microbiology results its severity. The way data were collected varied in these (n=3). Some hospitals reported that there was more than hospitals, with multiple systems being used in some of one mechanism of recording sepsis related information them. The most common mechanism was through clinical (Table 2.36). coding (n=21), adverse incident reporting systems (n=15),

Table 2.36 Hospital mechanism to centrally record all incidents of sepsis Recorded Yes % No % Subtotal Not Total answered Sepsis 46 21.2 171 78.8 217 9 226 Severe sepsis 43 19.8 174 80.2 217 9 226 Septic shock 46 22.0 163 78.0 209 17 226 Septicaemia 52 25.2 154 74.8 206 20 226

32 2

Key Findings

• 184/544 (33.8%) hospitals in this study had no formal • One in five hospitals (57/258; 22.1%) without critical sepsis protocol care facilities did not have formal arrangements for the • 309/343 (90.1%) hospitals with sepsis protocols had transfer of patients needing critical care based them on published guidelines • 55/215 (25.6%) acute hospitals utilised specialised • Most hospitals with protocols (305/321; 95%) stipulated proformas to identify and monitor patients with sepsis that action should be taken within one hour of diagnosis • 63/212 (29.7%) acute hospitals stated that there was of sepsis no policy in place covering staff handovers. However, • Of hospitals with protocols for recognition and 270/287 (94.1%) hospitals with a policy set aside time for management of sepsis, there was no formal education the formal handover of patients between doctors’ shifts in the use of the protocol on general wards for medical • The vast majority of acute hospitals (224/226; 99%) had staff in 65/305 (21.3%) and nursing staff in 86/314 an antimicrobial policy and although 139/204 (68.1%) (27.4%) of acute hospitals had daily microbiology ward rounds • In 518/532 (97.4%) hospitals, the hospital protocol on ICU (level 3), only 20/194 (10.3%) and 13/196 (6.6%) policies and guidelines were immediately available on of acute hospitals reported having daily microbiology the hospital intranet ward rounds on general medical or surgical wards (respectively). • The majority of hospitals without sepsis protocols (154/165; 93.3%) did have protocols for the • Only 29/519 (5.6%) hospitals in the study had leaflets to identification of the deteriorating patient give to patients to provide information about sepsis • 95/186 (51.1%) acute hospitals stated that there was • Only 78/215 (36.3%) acute hospitals had any form of a system in place for receiving a pre-alert for patients follow-up service for patients with sepsis arriving to the emergency department with sepsis • Half of the hospitals in the study (166/322; 51.6%) had • The vast majority (530/538; 98.5%) of hospitals have appointed a lead clinician for sepsis track and trigger systems for monitoring sick patients • Less than half of acute hospitals (90/204; 44%) were and these were uniformly linked to escalation protocols carrying out audit of the timely treatment of severe (516/527; 97.9%) sepsis • 199/223 (89.2%) hospitals with critical care facilities had • 43/217 (20%) hospitals had a means of centrally a Critical Care Outreach Team or equivalent and 96/196 recording incidents of severe sepsis (49%) of these were available 24/7

33 The organisation of care

34 3

Patient population and pre-hospital care Back to contents

Demographics Co-morbidities

The patient age distribution in the study was slightly older In this study 513 patients of a total of 569 (90%) had co- than the general population; about two-thirds of the morbidities on admission. This could be explained by the patients were above the age of 60. Male patients age distribution of patients included in the study and that comprised 56.2% of the study population (Figure 3.1). the inclusion criteria involved input from critical care. Figure Similar demographic patterns have been noted in other 3.2, overleaf, shows the top ten co-morbidities noted in this studies involving patients in critical care units with or study population, most of which are expected to predispose without sepsis.32 The distribution of Body Mass Index to infection and/or sepsis. Many patients had more than in the study group was also very similar to the general one co-morbidity and 254/569 (44.6%) had one or more population.33 of: diabetes mellitus, kidney disease and/or heart failure. The older age of the patient population in this study is the most likely explanation for the high prevalence of observed hypertension. The prevalence of hypertension in patients above the age of 65 is >50%.34

Number of patients Male Female

90

80

70

60

50

40

30

20

10

0 17-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 >90

Age (years)

Figure 3.1 Age and gender of the study population

35 Patient population and pre-hospital care

Number of patients (answers may be multiple)

250

200 44.8%

150 33% 26.1% 100

14.9% 50 14.3% 13.6% 12.8% 9.2% 9.2% 7.7% 0

Hypertension

Chronic lung disease Myocardial infarction Cerebrovascular disease Congestive heart failure Substance abuse/alcohol Peripheral vascular disease Diabetes mellitus type 1or 2 Permanent neurological deficit

Moderate or severe chronic kidney disease Figure 3.2 Top ten co-morbidities on admission (Clinician questionnaire n=513)

Smoking Previous admission with sepsis

Of the 541 patients in whom smoking history was available The treating clinicians reported a previous admission with 197 (36.4%) had never smoked. This is less than the sepsis in just over one quarter of the patients in this study. proportion in the general population (58%)34 (Table 3.1). Since we know that the diagnosis of sepsis can be missed and is often poorly documented, it is possible that more Table 3.1 Smoking history – Clinician questionnaire patients in our study had previously suffered an episode of Smoking history Number of % sepsis (Table 3.2). patients Current smoker 143 26.4 Table 3.2 Patient previously been admitted with sepsis – Clinician questionnaire Ex-smoker (<5 years) 68 12.6 Ex-smoker (>5 years) 133 24.6 Previous admission Number of % patients Never smoked 197 36.4 Yes 192 27.4 Subtotal 541 No 510 72.6 Unknown 151 Subtotal 702 Not answered 18 Insufficient data 8 Total 710 Total 710

Of the 192 patients who had documented evidence of sepsis in the past, the treating clinician was also able to note how long prior to the current admission the most recent admission occurred. Just over two-thirds had been admitted within the past year. Data were not available to estimate whether this may have been recurrence of the same infection or inadequately treated sepsis (Table 3.3).

36 3

Table 3.3 Length of time from previous admission – Table 3.5 Mode of admission to hospital – Clinician questionnaire Reviewers’ opinion Time from previous admission Number of % Mode of admission Number of % patients patients <1 month 39 22.5 Via the emergency department - 278 51.9 1-6 months 56 32.4 ambulance/air evacuation >6 months-1 year 19 11.0 Via the emergency department - 67 12.5 self referral > 1 year 59 34.1 Via the emergency department - 57 10.6 Subtotal 173 general practitioner referral Unknown 19 General practitioner referral - 44 8.2 Total 192 direct to ward Transfer from another hospital 27 5.0 Table 3.4 shows that a large majority of patients in this Elective admission 29 5.4 study (94.3%) were admitted as an emergency. Transferred from out-patients clinic 15 2.8 Via the emergency department - 8 1.5 Table 3.4 Type of admission – Reviewers’ opinion out of hours GP/111 call Admission type Number of % Transfer from psychiatric unit 4 0.7 patients Transfer from nursing home 4 0.7 Emergency 477 94.3 Via emergency department - other 3 0.6 Elective 29 5.7 Subtotal 536 Subtotal 506 Insufficient data 15 Unknown 45 Total 551 Total 551

Route of presentation/ admission to hospital Pre-hospital care

Reviewers looked at the mode of admission to hospital. General practitioner notes – peer reviewed by Over half of the admissions were brought to the hospital general practitioners emergency department (ED) by ambulance. Another fifth were referred directly by general practitioners (GPs) A total of 129 patients were identified though their hospital (including out-of-hours GP services) to the ED or acute case notes or on the clinician questionnaire that they had medical/ surgical units. Self-referral to a hospital ED formed seen their GP prior to admission. The relevant GP surgery was 12.5% of the total admissions (Table 3.5). Some hospital contacted and asked to provide all records of these patients’ admissions were not due to sepsis. These patients were previous 3 visits within the two-weeks prior to the admission included in the study because they developed sepsis whilst to hospital. A total of 54 sets of GP case notes were suitable in hospital. for review. They comprised 54 patient contacts immediately prior to hospital admission. In addition, 26 patients had notes for a second-to-last visit and 11 had third-to-last visit in the two-week period prior to hospital admission. The notes were anonymised and then peer reviewed by a panel of GPs as well as being used as part of the overall set of case notes reviewed by the wider panel of Reviewers.

37 Patient population and pre-hospital care

Number of patients GP in surgery Practice nurse in surgery GP home visit Other 30

25

20

15

10

5

0 Last visit before 2nd to last visit 3rd to last visit hospital admission

Figure 3.3 Type of visit to the surgery

Number of patients Home Nursing home Other

60

50

40

30

20

10

0 Last visit 2nd to last visit 3rd to last visit

Figure 3.4 Location immediately prior to general practitioner visit

Where the patient saw the GP prior to admission Location In this group of 54 patients seen prior to hospitalisation, half were seen at home, a quarter were seen in the surgery A large majority of patients were living at home prior to by the GP, and the remaining patients were consulted either admission, however 4/11 patients who required three visits on the phone or by an out-of-hours GP, practice nurse by the GP were in a nursing home compared with 3/54 telephone consultation or a nursing home visit (Figure 3.3). admissions at the first visit (Figure 3.4).

38 3

In the notes where data were available, 6 of 24 patients difficulty in making a diagnosis of early sepsis or judging its were seen by their GP on the day that they were admitted severity (Figure 3.5). to hospital. Data were missing in a number of cases. Most NCEPOD reports have highlighted the lack of adequate Telephone consultation documentation and it is seen here again that absence of precise times and dates has meant that the time of Where documentation of the telephone consultation was diagnosis of sepsis, and whether there was a delay was available, GP Reviewers checked whether appropriate advice not available in a number of cases. However, where it had been given. It was found that on their last visit patients was possible to establish timelines, delays were noted at usually received appropriate advice, possibly because sepsis different levels and on all visits, possibly emphasising the (and its severity) was easier to identify (Figure 3.6).

Number of patients Presenting to the GP surgery Referral to hospital GP recognition of the 12 severity of the illness 10

8

6

4

2

0 Last visit 2nd to last visit 3rd to last visit

Figure 3.5 Delay in presentation/diagnosis

Number of patients Yes No

18

16

14

12

10

8

6

4

2

0 Last visit 2nd to last visit 3rd to last visit

Figure 3.6 Appropriate advice given during the telephone consultation

39 Patient population and pre-hospital care

The main area of concern from GP Reviewers was around Assessments the adequacy of safety netting. Safety netting was only evident in 9/54 cases (last visit), 10/26 cases (2nd to last Most physiological vital signs were not recorded consistently visit) and 5/11 cases (3rd to last visit) (data not shown). by primary care teams, temperature and blood pressure Safety netting is a consultation technique used to ensure being noted in less than half of patients in this group. the timely re-appraisal of a patient whose status is uncertain Blood glucose were was recorded. It is possible that the or changeable. It should be part of any management plan GP making the assessment did not have access to the where uncertainty exists in the diagnosis, or where there equipment required to measure blood glucose. The vital may be a recognised risk of deterioration or complications signs were more consistently recorded in the final visit, as the disease process evolves over time. It involves possibly reflecting worsening clinical severity. Given that communicating clearly with the patient and their carers so these measurements form part of the diagnostic criteria for they understand the existence of uncertainty in diagnosis, sepsis they should be undertaken at each visit in acutely deterioration or complications. Red flag clinical features unwell patients to aid earlier diagnosis of sepsis (Tables 3.6 should be clearly explained with advice on when and how to and 3.7). seek help. Safety netting advice should also be documented in clinical notes.36

Tables 3.6 Assessment of vital signs and the healthcare provider that recorded them Last visit Assessment done Who made assessment Yes No Total GP Nurse Other Unknown Total Heart rate 33 21 54 31 1 1 0 33 Blood pressure 23 31 54 19 1 0 3 23 Respiratory rate 10 44 54 8 1 0 1 10 Temperature 25 29 54 18 1 0 6 25 Mental state 8 46 54 6 1 0 1 8 Blood glucose 2 52 54 0 0 2 0 2 Other 16 38 54 5 0 0 11 16

Tables 3.7 Documentation of vital signs at each of the three visits to the general practitioner prior to hospital admission Assessment Last visit (n/54) 2nd to last visit (n/26) 3rd to last visit (n/11) Heart rate 33 12 4 Blood pressure 23 8 2 Respiratory rate 10 4 4 Temperature 25 6 6 Mental state 8 3 1 Blood glucose 2 1 0 Other 16 3 5

40 3

Number of patients Low urine output Speech disturbance New rashes Confusion Severe muscle pain 8

7

6

5

4

3

2

1

0 Last visit 2nd to last visit 3rd to last visit

Figure 3.7 Symptoms of sepsis documented

Symptoms Diagnosis of sepsis missed/underestimated

GP case notes were also examined for documentation Table 3.8 summarises the situations where, in the GP of symptoms commonly associated with sepsis. As seen Reviewers’ opinion, the diagnosis or assessing the severity in Figure 3.7, only a small number of patients had these of sepsis could be improved. symptoms documented. Where recorded, some patients had more than one symptom of sepsis. New or acute confusion Early warning scores is an important feature in sepsis and should be sought along with all the others mentioned below. When looked No evidence was found in the case notes reviewed to for, confusion appeared most frequently (Figure 3.7). demonstrate that an early warning score had been used to assess and monitor physiological parameters. The GP Reviewers were able to see the benefit of such a score in a large majority of patients, where it could have helped in grading the sepsis severity (Table 3.9).

Table 3.8 Sepsis diagnosis missed by the general practitioner – GP Reviewers’ opinion Diagnosis of sepsis Severity of sepsis missed underestimated Last visit (n/54) 4 4 2nd to last visit (n/26) 4 1 3rd to last visit (n/11) 2 2

Table 3.9 An early warning score or track and trigger tool should have been used for this patient Yes No Subtotal Insufficient Total data Last visit 30 13 43 11 54 2nd to last visit 19 4 23 3 26 3rd to last visit 6 1 7 4 11

41 Patient population and pre-hospital care

Number of patients Yes No

30

25

20

15

10

5

0 Last visit 2nd to last visit 3rd to last visit

Figure 3.8 Adequacy of history taking – GP Reviewers’ opinion

History taking recording observations were also highlighted as missing. Reviewers did highlight good practice in communicating Reviewers assessed whether the history noted was adequate. and collaboration with patients and family members/carers Documentation was considered less than adequate in a in decision making. third to half of cases but improved at the last visit, possibly as the clinical picture of sepsis became clearer, or the patient Documentation of co-morbidities were more frequently deteriorated further (Figure 3.8). noted at the last visit before admission to hospital, which was appropriate. Patients displayed a wide spectrum of The themes emerging from comments made by Reviewers long-term conditions; predominantly cardiovascular and primarily related to absent documentation of symptoms respiratory (Table 3.10). like confusion, muscle aches and fever. Clinical signs and

Table 3.10 Co-morbidities documented by the general practitioner Co-morbidities Last visit 2nd to last visit 3rd to last visit Cardiovascular 29 15 0 Respiratory 16 1 0 Gastrointestinal 7 0 0 Neurological 11 2 0 Urological 3 0 0 Gynaecological 3 0 0 Renal 4 0 0 Oncological 6 0 0 Connective tissue disease 4 0 0 Diabetes mellitus 13 2 0 Other 18 3 6 None 7 3 5

42 3

Number of patients Yes No

40

35

30

25

20

15

10

5

0 Last visit 2nd to last visit 3rd to last visit

Figure 3.9 Treatment given

Given the importance of identifying the source of sepsis GP Reviewers were asked to give their opinion on the and instituting appropriate treatment, recognition and management of the patients at each visit (Figure 3.10 documentation of a possible source is essential. Reviewers overleaf). They were of the opinion that management was identified only 22 case notes where a provisional source appropriate in a larger proportion of patients at the last of sepsis was mentioned. Abdominal or pelvic infections visit when it was likely that the disease process had evolved. formed the largest group, followed by respiratory infection On the previous two visits, more patients received less and post-operative surgical causes (data not shown), than adequate management in terms of their assessment, although not all patients were commenced on treatment monitoring and treatment. Areas of deficiency included (Figure 3.9). the recording and monitoring of vital signs, maintaining adequate and legible records, delayed referrals, and relying When treatment was started, the type of treatment varied, on a telephone consultation, resulting in missed clues on with some patients receiving more than one modality of diagnosis and severity. A thorough clinical assessment therapy (Table 3.11). As was expected the most common should have been done in person. treatment initiated was antimicrobial therapy, since it is not always feasible to administer and monitor intravenous fluids or oxygen in general practice.

Table 3.11 Type of treatment given by the general practitioner Last visit 2nd to last visit 3rd to last visit Treatment given n/54 Appropriate n/26 Appropriate n/11 Appropriate IV fluids N/A N/A N/A N/A N/A N/A Oxygen 1 1 0 N/A 0 N/A Antimicrobials 12 9 14 8 6 5 Other 7 4 4 3 4 4

43 Patient population and pre-hospital care

Number of patients Yes No

45

40

35

30

25

20

15

10

5

0 Last visit 2nd to last visit 3rd to last visit

Figure 3.10 Appropriate management at the general practitioner visits – GP Reviewers’ opinion

Outcome of visit The details of the referral process are summarised in Table 3.13. The lack of information given to patients emerged as The majority of referrals to hospital were as an emergency a theme. The extent of safety netting was also a concern; after having multiple GP visits (Figure 3.11). Reviewers stated that in the majority of cases where it was not put in place that it should have been (Table 3.13). Reviewers felt that most referrals made were appropriate and timely, even when GPs may not have recognised the GP Reviewers were asked to provide their retrospective diagnosis or severity of sepsis in some patients (Table 3.12). opinion on whether management in primary care might

Number of patients Referred to hospital (emergency) Not referred to hospital Referred to hospital (non emergency) 35

30

25

20

15

10

5

0 Last visit 2nd to last visit 3rd to last visit

Figure 3.11 Outcome of visit to the general practitioner

44 3

Table 3.12 Appropriateness and timeliness of hospital referrals – GP Reviewers’ opinion Last visit (n/34) 2nd to last visit (n/2) Referral Yes No Yes No Appropriate 33 0 1 1 Timely 22 3 1 1 Successful 33 0 1 1

Table 3.13 Referral to hospital – GP Reviewers’ opinion Last visit 2nd to last visit 3rd to last visit (n/19) (n/24) (n/11) Yes No ID Yes No ID Yes No ID Correct decision 5 1 13 8 2 14 5 2 4 Information sheet 0 19 0 0 24 0 1 10 0 Safety netting 9 9 1 12 10 2 5 6 0 If ‘no’ – should there have been safety netting? 8 0 11 8 1 17 4 2 0 ID = insufficient data Table 3.14 Management in primary care affected have affected clinical outcome. In 23 of 54 patients the data the patient’s outcome – GP Reviewers’ opinion were inadequate to make a judgement. Of the remaining 31 Management affected outcome Number of patients, outcome was believed to be adversely affected in patients 12 patients. However, in two patients, outcome improved Yes 14 due to good practice (Table 3.14). No 17 Subtotal 31 On making a global assessment of care provided in primary Insufficient data 23 care, the GP Reviewers agreed that there was evidence of Total 54 good practice in 19 of the 54 case notes reviewed (Figure 3.12). Of the remainder, a large group (21) required organisational care. The reasons given for a less than good improvement in clinical care, nine required improvement grading included failure to assess and record vital signs and in organisation of care alone or both clinical and missed/delayed diagnosis of sepsis.

Number of patients

25

20

15

10

5

0 Good practice Room for Room for Room for Less than improvement improvement improvement satisfactory (clinical) (organisational) (clinical and organisational ) Figure 3.12 Overall quality of primary care – GP Reviewers’ opinion

45 Patient population and pre-hospital care

Primary care reviewed as part of the whole Table 3.15 Sepsis severity prior to admission – pathway from the secondary care perspective Reviewers’ opinion Sepsis status Number of % From the hospital case notes, Reviewers were asked to patients give their opinion on whether the patient had evidence of Infection only 98 32.9 infection only, sepsis, severe sepsis or septic shock as defined Sepsis 120 40.2 by the Surviving Sepsis Campaign sepsis screening tool1 Severe sepsis 61 20.5 prior to their hospital admission (Table 3.15). The Reviewers Septic shock 19 6.4 considered that in 298 patients there was evidence of infection and/or sepsis prior to admission to hospital. In this Subtotal 298 group, the Reviewers considered that 98/298 (33%) patients No evidence of infection prior to 73 had evidence of infection but not sepsis whilst 200 (67%) admission had sepsis, severe sepsis or septic shock. No available information on sepsis 180 status prior to admission Of those patients who were assessed in a pre-hospital Total 551 setting prior to admission, Reviewers found that the majority were seen by a GP or paramedic (Table 3.16) and Table 3.16 Healthcare professionals who assessed 52 patients were seen by both the GP and paramedics. the patient prior to admission ‘Other’ listed providers were primarily care home staff who Healthcare professionals Number of % accessed emergency services. Based on hospital case notes, patients Reviewers were of the opinion that the sepsis was missed Paramedic or equivalent 163 62.7 by GPs in 28/77 (36%) patients and in a quarter (19/72) the General practitioner 129 49.6 severity was underestimated (Table 3.17). Out of hours general practitioner/ 32 12.3 urgent care service The treating clinician at the hospital was asked whether Telephone consultation/111 11 4.2 there had been a delay in the patient presenting and being Other primary care provider 24 9.2 admitted to hospital. In their opinion, 112 patients (18%) Answers may be multiple n=260, not answered in 38 should have been admitted to hospital sooner (Table 3.18).

Table 3.17 Room for improvement in pre-hospital care - Reviewers’ opinion Room for improvement Yes No Subtotal Not Total answered Diagnosis missed by general practitioner 28 49 77 52 129 Severity underestimated by general practitioner 19 53 72 57 129 General practitioner missed the opportunity to refer 28 49 77 52 129

46 3

Table 3.18 Delay in the patient presenting to/being When these reasons for delay were compared with the admitted to hospital - treating clinicians’ opinion Reviewers’ opinion on reasons for delay in diagnosis of Delay Number of % sepsis in hospital inpatients, a similar theme of missed patients diagnosis appears. The diagnosis was missed in 55.4% Yes 112 18.0 patients with sepsis and up to 66% of patients with severe No 511 82.0 sepsis. The treating clinicians were of the opinion that 64/81 patients should have been admitted at least one day Subtotal 623 earlier (Table 3.20). These data highlight the need for better Unknown 53 awareness and use of tools for the diagnosis of sepsis. Not answered 34

Total 710 Table 3.20 Length of delay in arrival/admission to hospital Of the 112 patients considered to have experienced a delay Length of delay Number of % in presentation/admission to hospital, the treating clinicians patients were of the opinion that a large majority were due to 0-6 hours 9 11.1 patients seeking advice later than they should have. Delays >6-12 hours 6 7.4 from the GP (13) and EDs (11) were the next most common >12-24 hours 2 2.5 reasons, followed by the mental health of the patient (6), >1-2 days 30 37.0 the diagnosis being missed by staff in the nursing home (4) and lack of critical care beds (2) (Table 3.19). >2-4 days 16 19.8 >4-7 days 8 9.9 Table 3.19 Reason for delayed presentation/ >7-14 days 8 9.9 admission to hospital >14 days 2 2.5 Reason for delay Number of % Subtotal 81 patients Not answered 31 Patient did not seek medical help 66 59.5 early enough Total 112 General practitioner 13 11.7 Admitting hospital emergency 8 7.2 C A S E S T U D Y 1 department Other hospital emergency 3 2.7 A young patient presented to their GP with fever, department lethargy and dizziness. A diagnosis of viral infection Urgent care centre 3 2.7 was made. The following day, the patient deteriorated Community nurse 2 1.8 and called an emergency ambulance. On arrival, their 111 service 2 1.8 vital signs were recorded as pulse 124 bpm, BP 80/40 mmHg, respiratory rate 36/min and temperature 38.2 oC. Ambulance control centre 0 0.0 A diagnosis of severe sepsis due to community acquired Paramedic service 0 0.0 pneumonia was made following admission to hospital. Other 13 11.7 None of the above 1 0.9 The Reviewers felt that a standardised approach to Answers may be multiple n=111, not answered in 1 vital signs monitoring in primary care could have identified the low blood pressure at an earlier stage and helped to prevent deterioration.

47 Patient population and pre-hospital care

C A S E S T U D Y 2 Table 3.21 Pre-alert sent to the hospital Pre-alert sent Number of % A young patient developed fever with rigors, pain in patients their left flank and dysuria, but remained at home for Yes 8 9.4 4 days before going to see their GP. The patient was No 77 90.6 initially diagnosed with a flu-like syndrome but advised Subtotal 85 to return if they did not start to feel better. The patient re-presented to the GP surgery two days later and was Not applicable 4 noted “to be sleepy” and have a fever. A urine test was Insufficient data 44 suggestive of urinary tract infection and the patient was Total 133 transferred to hospital for admission. The infection was managed with intravenous antibiotics by the Critical Care early recognition and treatment, there appears to be scope Outreach Team and the patient recovered in 3 days. for further improvement in this area. In the organisational survey of hospitals 95/186 (51%) acute hospitals responded Reviewers were of the opinion that a urine test at that there was a system of receiving pre-alerts for sepsis. first presentation to the GP may have diagnosed the urinary tract infection. Appropriate treatment at that Vital signs and scores that use them (like the National Early stage might have prevented further deterioration, Warning Score9) help with early recognition of severe illness thus avoiding hospital admission. Drowsiness is an and can improve care, by providing a consistent way of important indicator of the severity of sepsis, which was communicating a patient’s condition as they move along a noticed on the second visit and resulted in a prompt pathway of care and are transferred between individuals, transfer to hospital. organisations and teams. Whilst most hospitals have moved to routine use of early warning scores calculated from recorded vital signs, ambulance and other services also regularly employ scales such as ‘alert, voice, pain, Of the 129 patients seen by a GP or out-of-hours GP, a unresponsive (AVPU)’ as a measure of level of alertness. pre-alert was sent to warn the emergency department or Details of vital signs recorded by primary and pre-hospital acute medical unit of the arrival of a patient with sepsis for care providers are shown in Table 3.22. Patients admitted by only eight patients (Table 3.21). Given the importance of the paramedics had the most comprehensive record.

Table 3.22 Details of pre-hospital vital signs Vital signs recorded GP % Paramedic % Other (n/129) (n/163) (n/24) Temperature 34 26.4 146 89.6 3 Blood pressure 32 24.8 157 96.3 5 Heart rate 40 31.0 163 100.0 6 Respiratory rate 8 6.2 159 97.5 2 Alert, voice, pain, unresponsive (AVPU) 8 6.2 144 88.3 1 Change in mental status 11 8.5 81 49.7 1 Blood glucose 3 2.3 129 79.1 1 Answers may be multiple

48 3

A total of 55/129 (43%) patients who saw their GP, or an Urgently starting treatment for sepsis is recommended. out-of-hours GP, arrived in hospital with a referral letter Reviewers found that 127/223 (57%) patients received some from their GP. Of these 55 letters, 34 were considered to treatment prior to arrival at hospital (Table 3.24) have all the relevant information (Table 3.23). Table 3.24 Pre-hospital treatment Table 3.23 Relevant information in referral letter from the general practitioner – GP Reviewers’ Treatment commenced Number of % opinion patients Relevant information Number of % Yes 127 57.0 patients No 96 43.0 Yes 34 63.0 Subtotal 223 No 20 37.0 Unknown 51 Subtotal 54 Not applicable 10 Insufficient data 1 Insufficient data 14 Total 55 Total 298

Table 3.25 Healthcare professional providing treatment Healthcare professional Fluids Oxygen Antimicrobials Other General practitioner (129) 0 0 19 6 Paramedic (163) 26 65 0 56 Other (2) 1 2 2 0 Inappropriate therapy 9 12 5 2

Table 3.26 Room for improvement in pre-hospital In those patients given pre-hospital treatment, intravenous care - Reviewers’ opinion fluids and oxygen were administered primarily by paramedics Room for improvement Number of % and GPs more commonly administered antimicrobials before patients referring patients to hospital as IV fluids and oxygen are rarely Yes 86 38.9 feasible to administer in general practice (Table 3.25). No 135 61.1 Subtotal 221 ‘Other’ treatments initiated by paramedics included Insufficient data 77 analgesia, nebulisers, glucose by mouth and anti-emetics. Inappropriate therapy included those cases where the Total 298 healthcare professional was unable to cannulate the patient or inadequate dosage of antimicrobials, fluids or oxygen Table 3.27 Details of room for improvement in pre- hospital care were administered. Area in need of improvement Number of % The Reviewers considered that there was room for patients improvement in pre-hospital management in 86/221 (39%) General practitioner 38 44.2 of patients sent to hospital (Table 3.26). Paramedic 41 47.7 Other primary care provider 4 4.7 The areas of improvement spanned all aspects of pre- Other 13 15.1 hospital care and often involved multiple steps in one Answers may be multiple n=86 patient’s journey (Table 3.27).

49 Patient population and pre-hospital care

Reviewers were of the opinion that the clinical outcome was The healthcare professional who assessed patients in the affected in six patients. The common reasons were delayed ED at triage and first review are summarised in Figure 3.13. recognition of sepsis or of its severity and the resultant delay Consultants were involved at presentation in a quarter of in timely intervention. patients but the majority of patients were reviewed by a specialist trainee or staff grade. Nurses were involved in Reviewers suggested that delays in recognition and triage in just over half the patients presenting with sepsis. management contributed to the patient’s deterioration and earlier intervention would have improved the patient’s C A S E S T U D Y 3 outcome. A young patient presented with a 2 day history The emergency department of cough and worsening shortness of breath on a background of bronchial asthma. The patient was A total of 369 patients were admitted through the hospital seen by their GP who diagnosed a chest infection and ED (Table 3.28) had sufficient notes to review the care in the transferred the patient by ambulance to the emergency ED. Table 3.29 shows the sepsis status of these patients at department. Reviewers were of the opinion that the this point in their care pathway according to the Reviewers; patient should have received intravenous fluids and the majority of patients were admitted with sepsis. The oxygen in the ambulance since they were manifesting Reviewers also assessed the documented diagnosis at this early hypotension and hypoxia. On arrival in the stage of the patient pathway and compared this with emergency department the triage nurse considered whether they considered the diagnosis to be correct. the possibility of chest infection but the hospital sepsis For infection, sepsis, severe sepsis and septic shock the proforma was not completed. Chest X-ray, antibiotics Reviewers were of the opinion that a number of cases were and initial assessment/management for sepsis was not not documented and hence possibly missed in each group. initiated. An hour later the patient became profoundly Only 10/91 patients had the diagnosis of severe sepsis hypotensive and drowsy at which time the patient was documented in their case notes (Table 3.29). reviewed by a consultant who then initiated the sepsis care bundle. The patient required transfer to critical care Table 3.28 Patients admitted via the emergency for mechanical ventilation and their condition improved department over the following 7 days. However, the ICU stay Emergency department Number of % was complicated by ventilator associated pneumonia admission patients and peripheral gangrene of both feet. There was Yes 369 75.0 significant disability at discharge requiring prolonged No 123 25.0 rehabilitation. Subtotal 492 Reviewers suggested that delays in recognition and Insufficient data 59 management contributed to the patient’s deterioration Total 551 and earlier intervention would have improved the patient’s outcome. Evaluation in most emergency departments normally involves initial assessment by a triage service and then an assessment by a senior clinician. Reviewers considered that there was a delay in initial triage in 27/294 (9.2%) patients, and in 112/279 (40%) patients there was a delay in senior review (Table 3.30). Eighteen of the patients delayed in triage were also delayed for senior review and 16 of these had sepsis, severe sepsis or septic shock at this time-point.

50 3

Table 3.29 Sepsis status in the emergency department Sepsis status in the emergency Reviewer opinion % of emergency Number of patients in department - number of department each group documented patients population in the case notes (%) Infection only 48 14.0 30 (62.5) Sepsis 142 41.4 76 (53.5) Severe sepsis 91 26.5 10 (11.0) Septic shock 45 13.1 15 (33.3) None of the above 17 5.0 15 (88.2) Subtotal 343 Insufficient data 26 Total 369

Table 3.30 Delayed review in the emergency department Delay In triage % In senior review % - number of patients - number of patients Yes 27 9.2 112 40.1 No 267 90.8 167 59.9 Subtotal 294 279 Insufficient data 75 90 Total 369 369

Percentage Triage Review

60

50

40

30

20

10

0

Other

Consultant Staff grade Basic grade Staff nurse

Senior staff nurse

Senior specialist traineeJunior specialist trainee

Nurse consultant/practitioner

Figure 3.13 Grade of clinician who assessed the patient in the emergency department

51 Patient population and pre-hospital care

Table 3.31 Appropriateness of the grade of clinician who assessed the patient in the emergency department Appropriate Triage % Senior % grade grade Yes 168 96.0 161 89.4 No 7 4.0 19 10.6 Subtotal 175 180 Insufficient data 194 189 Total 369 369

Reviewers were of the opinion that an appropriate grade Many previous NCEPOD reports have commented on the of staff triaged the patient in 168/175 (96%) cases and importance of documentation yet it is an area that always 161/180 (90%) patients were reviewed in the emergency needs improvement. department by a clinician of appropriate seniority (Table 3.31). Confusion and delirium can be important indicators of developing or worsening sepsis and are easier to measure Having a clear and consistent record of vital signs is at the bedside compared to some vital signs that require essential. It was found that none of the vital signs were an instrument. The common scores in use are ‘AVPU’ or the recorded consistently in the notes. Given the importance of Glasgow Coma Scale. Despite their ease, they were recorded level of consciousness and mental status in the diagnosis of in only 69.4% of patients (Table 3.32). sepsis, this area needs improvement. It is known that identifying the source of sepsis and At triage there were 66 patients for whom none of the controlling it is vital in managing sepsis.2 Reviewers noted listed vital signs were recorded and 150 patients at the that there were 173 cases at triage and 64 at the ED review, stage of ED senior review, and 37 cases where none of the in whom the likely source of infection was not documented listed vital signs were recorded at either assessment. At (Table 3.33). In half of the cases not documented at triage triage, 103 sets of case notes had all the listed vital signs (59/120) and one third of those not documented at the ED recorded. For 152 cases there were a complete set of vital review (17/51) the Reviewers felt that it should have been. signs recorded between the two assessments (Table 3.32).

Table 3.32 Vital signs recorded in the emergency department Vital signs At triage - number % of cases assessed Either at triage % of cases assessed recorded of patients at triage or emergency either at triage department review or emergency - number of department review patients Temperature 279 75.6 306 82.9 Blood pressure 296 80.2 326 88.3 Heart rate 296 80.2 327 88.6 Respiratory rate 283 76.7 310 84.0 GCS/AVPU/mental 217 58.8 256 69.4 status Blood glucose 145 39.3 167 45.3 Answers may be multiple n=369

52 3

Percentage % of cases assessed % of cases assessed either at triage at triage or emergency 100 department senior review

90

80

70

60

50

40

30

20

10

0 Temperature BP Heart rate Respiratory rate GCS/AVPU/ Blood glucose mental status Figure 3.14 Vital signs assessed in the emergency department n=369

The likely source of infection/sepsis was documented at triage For cases where the source of infection was not in 148/321 (46%) patients, which increased to 227/291 documented, Reviewers were of the opinion that there was (78%) patients at senior review. It is important to consider adequate evidence for a provisional diagnosis of the source the possible, or most likely, source of infection as it helps in of infection in 59/120 patients at triage and 34/151 patients selecting the most appropriate antimicrobial and any other at senior review (Table 3.34). intervention required to treat the source of infection.

Table 3.33 Likely source of infection was documented Likely source of infection was documented Triage % Review % Yes 148 46.1 227 78.0 No 173 53.9 64 22.0 Subtotal 321 291 Insufficient data 48 78 Total 369 369

Table 3.34 If not documented, the likely source of infection should have been documented - Reviewers’ opinion Source should have been documented Triage % Review % Yes 59 49.2 34 66.7 No 61 50.8 17 33.3 Subtotal 120 51 Insufficient data 53 13 Total 173 64

53 Patient population and pre-hospital care

The Reviewers considered the quality and completeness of Table 3.36 Detail of missing information from the initial assessment process and stated that the history treatment and monitoring plans was often well documented, but that there was room for Treatment plan Number of % improvement in the investigations (95/369; 26%), treatment patients plans (117/369; 32%) and monitoring plans (136/369; 37%) (n=117) (Table 3.35). The detail of missing information from the Oxygen 59 50.4 treatment and monitoring plans are shown in Table 3.36. Fluids 39 33.3 Antibiotics 44 37.6 Table 3.35 Areas needing improvement in the initial assessment Others 21 17.9 Monitoring plan Number of % Room for improvement Number of % patients patients (n=136) History taking 4 1.1 Urine output 106 77.9 Investigations 95 25.7 Early warning score 75 55.1 Treatment plan 117 31.7 Other 20 14.7 Monitoring plan 136 36.9 Answers may be multiple Answers may be multiple n=369

There were eight patients for whom the treatment plan in 37% cases (data not shown) and were inappropriately was missing for all three interventions: oxygen, fluids and delayed in another 9.5%. Where other cultures were required antimicrobial therapy (Table 3.36). Prompt cultures, which 11.8% of these were incomplete. In the area of diagnostic are a key part of diagnosis and treatment, were missing imaging 20.6% CT scans were delayed (Figure 3.15).

Percentage Carried out Delayed Incomplete

100

90

80

70

60

50

40

30

20

10

0 CT LF Ts U&E Amylase Glucose Lactate Chest xray Blood gases Other tissue/ Blood cultures fluid culture Full blood count Coagulation screen Figure 3.15 The percentage of cases for whom the listed investigations were carried out in the emergency department

54 3

Key Findings

• One quarter of patients (192/702; 27.4%) in this study • Deficiencies in record keeping were present in both had been admitted previously with an episode of sepsis primary and secondary care • 1 in 8 patients (67/536; 12.5%) self-referred to hospital • The commonest reason for delay in arriving at the hospital with sepsis emergency department was because the patient did not • In only 8/85 (9.4%) patients seen by a GP (where the present to a clinician early enough (66/111; 59.4%) reviewers could answer) were pre-alerts sent to warn • 267/294 (90.8%) patients admitted via the emergency hospitals of the arrival of a patient with sepsis department had appropriately timed triage assessment • Both the secondary care Reviewers assessing hospital • 112/279 (40.1%) patients did not have a timely review case notes and the GP Reviewers reviewing the GP notes by a senior clinician found that there was a poor adherence to the recording • There was inconsistency in the recording of vital signs in of vital signs by GPs assessing patients. Less than half of the emergency department with 66/369 (17.8%) having patients had their temperature (25/54) or blood pressure no vital signs recorded in their case notes (23/54) taken • A possible source of infection was only recorded at • Evidence of safety netting was present in 9/54 cases (last triage in 148/321 (46.1%) of patients admitted via the visit), 10/26 cases (2nd to last visit) and 5/11 cases (3rd emergency department to last visit) • Reviewers considered that there was room for • For just over half (55/101) of the patients referred to improvement in the emergency department in hospital from the GP, the referral letter was included in investigations (95/369; 25.7%), treatment planning the case note record (117/369; 31.7%) and monitoring plan (136/369; • There was room for improvement in 86/221 (38.9%) in 36.9%) the care provided to patients in the primary care setting • No early warning score was used in any of the GP case notes reviewed

55 Admission

56 4

Admission to hospital Back to contents

There was no difference seen in the days of the week on However, slightly more patients were admitted out of which patients in the study were admitted to hospital. normal working hours (Figures 4.1 and 4.2).

Percentage

20

18

16

14

12

10

8

6

4

2

0 Sunday Monday Tuesday Wednesday Thursday Friday Saturday

Figure 4.1 Day of admission to hospital

Percentage

45

40

35

30

25

20

15

10

5

0 Monday to Friday Monday to Friday Weekend 08.00-17.59 18.00-7.59 Figure 4.2 Time of admission to hospital

57 Admission to hospital

Table 4.1 Location to which the patient was first Following assessment and triage, Table 4.1 shows the admitted hospital location to which patients were first admitted. Location Number of % patients Figure 4.3 demonstrates that as severity of sepsis increased Acute admissions unit 275 53.1 the proportion of patients admitted to higher care areas General ward 72 13.9 also increased. This is entirely in line with what would and Specialist ward 75 14.5 should be expected. Level 2 42 8.1 One in five patients were admitted to higher care areas. The Level 3 44 8.5 Reviewers were of the opinion that in 93% (493/530) of Theatre 10 1.9 cases this was the correct decision. Of the 37/530 patients Subtotal 518 considered to have been admitted to an incorrect area, Insufficient data 33 27 should have been admitted to a critical care facility for Total 551 higher level care (Tables 4.2 and 4.3).

Percentage of patients

70

60

50

40

30

20

10

0 No infection Infection only Sepsis Severe sepsis Septic shock

AMU/SAU/CDU % General ward % Specialist ward % HDU % ICU % Theatre %

Figure 4.3 Location admitted to and severity of sepsis

Table 4.2 Correct location for the point admission – Table 4.3 If the patient was admitted to an Reviewers’ opinion inappropriate location ­– level of care to which the patient should have been admitted to? Correct location Number of % patients Level of care Number of Yes 493 93.0 patients No 37 7.0 Level 0/1 9 Subtotal 530 Level 2 26 Insufficient data 21 Level 3 1 Total 551 Subtotal 36 Insufficient data 1 Total 37

58 4

The Reviewers considered that 49/361 (13.6%) patients There were 80/466 (18%) patients who were not reviewed experienced a delay in their admission to a hospital bed by a consultant within an adequate time frame in the view (Table 4.4). of the Reviewers (Table 4.7). Royal College of Physicians of London guidelines state that all emergency admissions Table 4.4 Admission to the ward was delayed – must be seen and have a thorough clinical assessment by Reviewers’ opinion a suitable consultant as soon as possible but at the latest Admission delayed Number of % within 14 hours from the time of arrival at hospital. This is patients also the standard set by NHS England on 7 day services.37,38 Yes 49 13.6 No 312 86.4 Table 4.7 Timeliness of first consultant review – Subtotal 361 Reviewers’ opinion Insufficient data 190 Timely consultant review Number of % Total 551 patients Yes 366 82.1 The principal reason for the delay to admit to a ward is shown No 80 17.9 in Table 4.5. In 16/30 cases this was due to a lack of beds. Subtotal 446 Table 4.5 Reason for delay to admit to a ward – Insufficient data 105 Reviewers’ opinion Total 551 Reason admission to ward was Number of delayed patients Assessing the time frame from admission to consultant Lack of beds 16 review 20% (116/571) of patients were reviewed by a Delays in the emergency department 8 consultant more than 14 hours after admission (Table 4.8). Portering delay 2 Clinical reason 2 Table 4.8 Time from admission to consultant review – Clinician questionnaire Delayed investigations 1 Time from admission to Number of % Lack of staff 1 consultant review patients Subtotal 30 Negative time - seen in the 42 7.4 Not answered 19 emergency department before Total 49 admission 0-1 hours 80 14.0 The Reviewers were of the opinion that the delay in admission >1-4 hours 107 18.7 to the ward affected the outcome in 7/37 patients (Table 4.6), >4-6 hours 69 12.1 four of whom died. >6-8 hours 48 8.4

Table 4.6 Outcome affected by the delay in >8-12 hours 88 15.4 admission to hospital – Reviewers’ opinion >12-14 hours 21 3.7 Outcome Number of >14-24 hours 86 15.1 patients >24 hours 30 5.3 Yes 7 Subtotal 571 No 30 Missing data 139 Subtotal 37 Total 710 Insufficient data 12 Total 49

59 Admission to hospital

Focusing on the data for those patients who arrived at Table 4.10 Changes made to patient care following hospital with sepsis (filtering out those that developed sepsis a consultant review – Reviewers’ opinion in hospital), the proportion of patients seen by a consultant Changes made Number of % after 14 hours is the same (95/471; 20%). patients Diagnosis of sepsis 38 13.5 Of those patients who had a delayed (more than 14 hours) Documentation of diagnosis of 34 12.1 consultant review, 65/116 (56%) were admitted out-of- sepsis hours or on the weekend. Only 9/116 (8%) patients in this Documentation of severity of 22 7.8 group were admitted directly to critical care. sepsis Investigations 154 54.8 Successive NCEPOD reports have highlighted the benefits Other 62 22.1 of consultant review for acutely unwell patients, as has Treatment plan 185 65.8 the recent Ombudsman report into sepsis.21 In this study a consultant did not review 18% of patients in a timely Starting care bundle 7 2.5 manner. In view of the fact that changes were made to the Monitoring plan 76 27.0 patients’ management in 62% of cases following consultant Answers may be multiple n=281 review (Table 4.9) and following this review 66% of these patients had a change in the treatment plan, it remains C A S E S T U D Y 4 important that patients receive prompt consultant input. An elderly patient was admitted to a small district Table 4.9 Changes made following the first general hospital with abdominal pain and vomiting. The consultant review – Reviewers’ opinion patient was diagnosed with gallstone pancreatitis and Changes made Number of % was given antibiotics and supportive treatment. The patients inpatient notes for the admission were poor and there Yes 281 61.5 was no evidence of senior input. After two weeks the No 176 38.5 patient was transferred to a tertiary unit with a necrotic pancreas for percutaneous drainage of a peripancreatic Subtotal 457 collection. Over the next two weeks the patient’s Insufficient data 94 condition slowly deteriorated and the patient died. Total 551 The Reviewers were of the opinion that there had The principal changes made by consultants were to the been inadequate senior review, there was no clear investigations ordered and the treatment plans (Table 4.10). management plan and that initial fluid resuscitation had Consultant involvement for patients considered ‘high risk’ been inadequate. Earlier structured treatment may have (with a high risk of mortality, or where a patient is unstable produced a better outcome. and not responding to treatment as expected) should be within one hour. Numerous reviews have concluded that patients have increased morbidity and mortality when there is a delay in the involvement in their care of consultants. This has been seen across a wide range of specialties including in acute medicine and acute surgery, emergency medicine, trauma, anaesthetics and obstetrics.39

60 4

Key Findings

• In the Reviewers’ opinion, 493/530 (93.0%) patients were admitted to the correct location • In the Reviewers’ opinion, 49/361 (13.6%) of patients were delayed in their admission to a definitive hospital bed. The principal reason for delay (16/30) was a lack of beds • The delay in admission to hospital affected the outcome in 7/37 patients • 80/446 (17.9%) patients were not reviewed by a consultant within an adequate time frame according to Reviewers • 116/571 (20.3%) patients in this study were not seen by a consultant within 14 hours, even for those who arrived in hospital with sepsis (95/471; 20%) • 281/457 (61.5%) patients had changes made to their care following consultant review

61 Diagnostic pathway

62 5

Patients with hospital-acquired infections Back to contents

The Health Protection Agency identified the prevalence of Table 5.2 Source of the hospital-acquired infection hospital-acquired infections to be 6.4% in 2011 compared Source of infection Number of to 8.2% in 2006.40 The most frequent hospital-acquired patients infections detected were respiratory tract, urinary tract and Chest infection/hospital-acquired 38 surgical site infections. The prevalence of hospital-acquired pneumonia/aspiration pneumonia infections and device use was highest in intensive care units, Directly related to a procedure 30 which relates in part to the complexity and vulnerability of Related to the post-op care following a 10 patients in this setting. In the current study, according to procedure the Reviewers 115/498 patients developed their infection Catheter 8 in hospital (Table 5.1). Hospital-acquired infections of the Ventilator acquired pneumonia 5 respiratory tract and relating (either directly or indirectly) to Gastrointestinal 4 a surgical procedure formed the majority of cases, where it Cellulitis 3 was possible to identify the source of infection (Table 5.2). Infected pressure sores 2 Table 5.1 Patient developed the infection that Cannula 2 caused the episode of sepsis whilst in hospital – Urinary 2 Reviewers’ opinion Norovirus 2 Hospital-acquired infection Number of % Intravascular device 1 patients Insufficient data 20 Yes 115 23.1 Answers may be multiple n=115 No 383 76.9

Subtotal 498 Table 5.3 Infection occurred following an invasive Not answered 53 procedure Total 551 Infection Number of % patients Yes 73 63.5 More than half of these patients (73/115; 63%) appear No 42 36.5 to have acquired their infection following an invasive procedure (Table 5.3). Total 115

The commonest procedure resulting in sepsis was a Table 5.4 Prophylactic use of antimicrobials procedure involving the abdomen (29 cases). In 15 cases Prophylactic antimicrobials Number of % where antimicrobials were not given prophylactically, the given patients Reviewers considered that they should have been in only Yes 43 74.1 one case (Table 5.4). No 15 25.9 Subtotal 58 Insufficient data 15 Total 73

63 Patients with hospital-acquired infections

It has been demonstrated that increased compliance with Table 5.6 Time from surgery to infection sepsis performance bundles is associated with better Days Number of % outcomes, and this must include attempts at prevention as patients well as cure.41 In this study there was no evidence in nine Same day 8 11.9 patients out of 39 that a surgical site bundle was used 1-2 26 38.8 (Table 5.5). 3-7 19 28.4 Table 5.5 Evidence that a surgical site bundle was 8-14 9 13.4 used 15-21 1 1.5 Evidence of surgical site bundle Number of 22-30 2 3.0 patients >30 2 3.0 Yes 30 Subtotal 67 No 9 Missing data 6 Subtotal 39 Total 73 Insufficient data 34 Total 73 Table 5.7 Preventable hospital-acquired infection – Reviewers’ opinion

The majority of surgical site infections are preventable. Preventable infection Number of % patients Measures may be taken throughout the surgical pathway to reduce the risk of infection. NICE Clinical Guideline Yes 10 11.4 74 makes recommendations for the prevention and No 78 88.6 management of surgical infections based on best available Subtotal 88 published evidence and the use of a bundle of care.42 Insufficient data 27 Total 115 It can be seen in Table 5.6 that for those patients who developed infection/sepsis following surgery, sepsis was first Table 5.8 shows the details of themes of these cases. It recorded within one week of the procedure, in the majority appears that most of these cases were failures in basic care of cases. and recognition of a deteriorating patient.

The Reviewers considered that in 10/88 patients who The Reviewers found evidence in the majority of cases acquired an infection in hospital that it might have been (71/83) with hospital-acquired infections that patients preventable (Table 5.7). were monitored using an early warning scoring system (Table 5.9). The Reviewers considered that the frequency

Table 5.8 Details of ‘preventable’ hospital-acquired infections “Possibly preventable as the patient had come in from a nursing home with pressure sores - developed gangrenous pressure sores in hospital. No mention of specialist mattresses etc being used on the acute ward.” “There was an unusual plan of initial prolonged post-operative bed rest which probably contributed to post-operative pneumonia.” “Long line inserted using 'aseptic technique' but should have been a sterile procedure.” “There is no evidence that I can see that pre-op/ intra-op antibiotics were administered - this is unusual I would say for a patient having an emergency laparotomy.” “Possibly preventable by pre-op MSU screening. The organism isolated later likely to be resistant to prophylaxis used.”

64 5

of monitoring during the 24 hours before sepsis was Table 5.10 Adequacy of the frequency of monitoring documented was appropriate in 90% (69/77) of patients during the 24 hours in hospital prior to sepsis being who developed a hospital-acquired infection (Table 5.10). first documented - Reviewers’ opinion Adequate monitoring Number of % Table 5.9 Use of early warning scores in patients patients who developed hospital-acquired infections Yes 69 89.6 Early warning score used Number of % No 8 10.4 patients Subtotal 77 Yes 71 85.5 Insufficient data 38 No 12 14.5 Total 115 Subtotal 83 Insufficient data 32 Total 115 adequate; where it was considered to be inappropriate, the Reviewers were of the opinion that for a patient with sepsis, NICE Clinical Guideline 50, on the recognition and response the monitoring should have occurred more frequently than to acute illness in adults in hospital states that, “physiology the standard 6 or 12 hourly observations. track and trigger systems should be used to monitor all adult patients in acute hospital settings. Physiological The Reviewers considered that 23/95 patients who acquired observations should be monitored at least every 12 hours. an infection in hospital could have been identified sooner, The frequency of monitoring should increase if abnormal and 24/95 should have been treated in a more appropriate physiology is detected”.47 In the current study, plotting the manner. In the group considered by Reviewers to have frequency of monitoring against the Reviewers’ opinion on been treated inappropriately, the commonest reason for its appropriateness (Figure 5.1), it can be seen that although being classified as such was the choice of an inappropriate the frequency of monitoring was considered to be largely antimicrobial therapy and/or an inappropriate dose.

Number of patients Frequency monitoring Frequency monitoring appropriate inappropriate 20

18

16

14

12

10

8

6

4

2

0 Hourly or more Every >1-2 Every >2-4 Every >4-6 Every >6-12 >12 hourly frequent hours hours hours hours

Figure 5.1 Frequency and appropriateness of monitoring in the 24 hours before sepsis was diagnosed

65 Patients with hospital-acquired infections

Key Findings

• 115/498 (23%) patients acquired their infection whilst in hospital. In half of these patients 73/115 (63.5%) the infection was diagnosed following an invasive procedure • A surgical site bundle was utilised in 30/73 (41.1%) invasive procedures • In 10/88 (11.4%) patients with hospital-acquired infection, the Reviewers stated that the infection was preventable • 23/95 (24.2%) patients could have had their infection identified sooner and 23/63 (36.8%) should have commenced treatment sooner in the opinion of Reviewers

66 6

First identification of sepsis Back to contents

The systemic inflammatory response syndrome (SIRS) was Table 6.1 Positive SIRS criteria and organ first described 23 years ago as a clinical expression of the dysfunction at time of sepsis diagnosis – Clinician questionnaire immune system’s response to an infection. In the presence of symptoms meeting two or more of the SIRS criteria, Positive SIRS criteria Number of % severe sepsis was seen as evolving from infection to sepsis, patients severe sepsis, and septic shock, in order of increasing Tachycardia (>90bpm) 513 72.3 severity. However, the need for patients to meet two or Leukocytosis (WBC>12000) 446 62.8 more SIRS criteria has been widely criticised because of a Tachypnoea (>20bpm) 415 58.5 low specificity for infection within 24 hours after admission. Hyperthermia (>38.3oC) 324 45.6 Moreover, some patients (the elderly and those taking Measures of organ dysfunction medications that affect heart rate, respiratory rate, or body Lactate (>2mmol/L) 271 38.2 temperature) may not have symptoms meeting two or more Altered mental status 246 34.6 SIRS criteria, despite having infection and organ failure.43 The incidence of SIRS criteria and organ dysfunction at the Systolic BP <90mmHg/MAP 230 32.4 <65mmHg time of sepsis diagnosis, identified from the patient’s notes is listed in Table 6.1. Creatinine (>178µMol/L) 165 23.2 Hypothermia 101 14.2 The Reviewers considered that there had been a delay in Platelet count (<100000µl) 98 13.8 identifying sepsis in 182/505 (36.0%) patients. There was a Coagulopathy (INR>1.5) 91 12.8 delay in identifying severe sepsis in half (167/324; 51.5%) Hyperglycaemia (plasma glucose 90 12.7 the patients and in one third of those with septic shock >7.7mmol/L) (63/193; 32.6%) (Figure 6.1 overleaf). Bilirubin (>34.2mmol/L) 87 12.3 Acute lung injury 82 11.5 The median delay in identifying sepsis was 9 hours and the Systolic BP decrease (>40mmHg 63 8.9 mode 3 hours. A graph of the length of delay in identifying from baseline) sepsis, severe sepsis and septic shock and the cumulative Leukopaenia (WBC <4000) 33 4.6 percentage of patients is seen in Figure 6.2 overleaf. Answers may be multiple n=710 These data may add support to the difficulty healthcare professionals may encounter in identifying sepsis, and it is only as it becomes severe and moves to septic shock that diagnosis is made easier.

67 First identification of sepsis

Percentage No delay in identification Delay in identification

100

90

80

70

60

50

40

30

20

10

0 Sepsis n=505 Severe sepsis n=324 Septic shock n=193

Figure 6.1 Delay in identifying sepsis, severe sepsis and septic shock – Reviewers’ opinion

Cumulative percentage Sepsis Severe sepsis Septic shock

100

90

80

70

60

50

40

30

20

10

0 0 10 20 30 40 50 60 70 80 90 100 Time (hours)

Figure 6.2 Cumulative percentage of time delay in diagnosing sepsis, severe sepsis and septic shock

68 6

Percentage Delay identifying sepsis No delay identifying sepsis

90

80

70

60

50

40

30

20

10

0 Good practice Room for Room for Room for Less than n=171 improvement improvement improvement satisfactory (clinical) (organisational) (clinical and n=30 n=127 n=32 organisational) n=102

Figure 6.3 Overall quality of care and delay in diagnosis

Delay in identifying sepsis and the Reviewers’ opinion on diagnosis increases. It is not surprising that the delay in overall quality of care seemed to be linked (Figure 6.3). As diagnosis would be a key contributor to the quality of care the rated quality of care decreases with recognised room for rating as a delay in diagnosis of sepsis leads on to a delay in improvement, the proportion of each group with a delayed all subsequent management steps.

Table 6.2 Reason for the delay in diagnosis of sepsis Reason for delay in diagnosis: Sepsis: % Severe % Septic % Number of sepsis: shock: patients Number of Number of patients patients Incorrect calculation of early warning score 3 1.7 1 0.6 0 0 Missed by reviewing clinician 97 55.4 105 66.0 36 62.1 Lack of senior review 18 10.3 10 6.3 6 10.3 Insufficient frequency of clinical review 7 4.0 5 3.1 4 6.9 Insufficient monitoring/investigations 6 3.4 9 5.7 0 0 Other 44 25.1 29 18.2 12 20.7 Subtotal 175 159 58 Not answered 7 8 5 Total 182 167 63

69 First identification of sepsis

The reasons for delay in diagnosis are summarised in Table C A S E S T U D Y 5 6.2. The most common reason for delay in diagnosis of sepsis, severe sepsis and septic shock in the Reviewer’s An elderly patient with type 2 diabetes and opinion was that clinicians were missing the diagnosis hypertension presented to the emergency department on review. Clinicians should be more alert to the possible with a two-day history of hip pain. They were diagnosis of sepsis in unwell patients. Senior clinical input discharged home with anti-inflammatory medication. has an important part to play as does increasing awareness Two days later the patient presented again with hip through training and education of staff. There may also be pain and generalised body pain. They were referred to an argument for the wider use of tests/markers of sepsis such the surgeons as they had some abdominal pain and a as procalcitonin. In the current study, procalcitonin was only lactate of 9.4. The surgeon organised an abdominal used in 11/710 (1.5%) of patients (data not shown). The ultrasound and referred the patient to the physicians procalcitonin test is relatively new, but it is beginning to be suggesting a diagnosis of arthritis and possible sepsis. used in UK hospitals. Recent studies have shown that it may The junior physician considered sepsis or vasculitis as a help to discover whether a seriously ill person is developing diagnosis. Senior review shortly after this, resulted in a sepsis. It has been studied mainly in emergency departments diagnosis of severe sepsis secondary to septic arthritis. or critical care in patients who have symptoms that may be The patient was transferred to a high dependency unit due to sepsis. For diagnosis, procalcitonin is best used on the and developed multiple-organ failure, requiring renal first day the patient is seen. It may be used later on to follow support. A long inpatient stay followed but they were how the patient responds to treatment. However, there is discharged home with difficulty walking requiring important evidence that having robust systems in place to extensive rehabilitation. increase awareness, support the decision making process and expedite correct treatment, which encompass all of the above The Reviewers were of the opinion that this case is key to driving improvement. A formal bundle/pathway for demonstrated that the identification of sepsis can be sepsis that includes a screening tool for identifying patients difficult despite being seen by multiple specialties. This and an early warning scoring system linked to escalation case also demonstrates the value of early senior input protocols for the management of the deteriorating patient for acutely ill patients. would be a part of such a system.

Table 6.3 Use of sepsis screening tool to diagnose According to the Reviewers, 128/479 (26.7%) patients with sepsis sepsis were identified using formalised screening tools or Diagnosis made using Number of % scoring systems designed to detect sepsis or physiological patients deterioration (Table 6.3). Sepsis screening tool 62 12.9 Other track & trigger tool 15 3.1 However, according to the treating clinicians completing National early warning score 51 10.6 the questionnaire, 344/580 (59%) were monitored on early (NEWS) warning scoring systems at the time of sepsis identification None of the above - clinical signs 351 73.3 (data not shown); the difference is likely to reflect issues only with documentation. Subtotal 479 Insufficient data 72 Total 551

70 6

Percentage Delay in diagnosis No delay in diagnosis

100

90

80

70

60

50

40

30

20

10

0 Tool used None-clinical signs Tool used None-clinical signs Tool used None-clinical signs Sepsis n=505 Severe sepsis n=324 Septic shock n=193

Figure 6.4 Effect of using a screening tool on the delay in diagnosis of sepsis

Figure 6.4 explores the relationship between the use of In common with previous NCEPOD reports the Reviewers screening tools/early warning scores and delay, in the commented on the poor standard of documentation Reviewers’ opinion, for the identification of sepsis, severe with over 70% of the dataset having less than good sepsis and septic shock. In those patients with severe sepsis documentation of sepsis (Table 6.4). Lack of documentation there appeared to be an association between failure to use has been previously shown to be associated with poor a screening tool/ early warning score and delay in making a outcomes,44 and clear clinical pathways are associated diagnosis. However, there appeared to be no difference in with reduced in-hospital complications and improved the identification of septic shock. This could be because the documentation without negatively impacting on length symptoms of septic shock are more obvious to the reviewing of stay.45 It was of note therefore, that in those patients in clinician. It should be noted that even when a screening whom the Reviewers considered the documentation of sepsis tool/ early warning score was employed there was a delay to be good, there was less delay in the diagnosis of sepsis. in diagnosing 35% of the severe sepsis cases. The use of screening tools or early warning scores would be just one Table 6.4 Documentation of ‘sepsis’ in the case of a series of measures to help improve the sepsis pathway notes – Reviewers’ opinion of care. Sepsis documentation Number of % patients Clinical algorithms have proved useful in expediting the Good 152 29.0 diagnosis (and subsequent management) of a number of Adequate 212 40.5 acute conditions (for example, in the management of acute Poor 160 30.5 chest pain) and the use of a sepsis bundle that includes a Subtotal 524 screening tool may be beneficial. The impact of a sepsis Insufficient data 27 bundle on management of patients with sepsis will be explored in more detail in the next section. Total 551

71 First identification of sepsis

C A S E S T U D Y 6 Table 6.6 Detail of vital signs measured and recorded A patient developed a hospital-acquired pneumonia Vital signs taken Number of % following a laparotomy for bowel obstruction. Although patients there was prompt identification of the pneumonia, key GCS/AVPU 218 52.9 microbiological investigations and arterial blood gases Temperature 378 91.7 were omitted. There was also no mention of triggering Blood pressure 364 88.3 the sepsis pathway or utilising a care bundle. Sepsis was Heart rate 387 93.9 diagnosed by the Critical Care Outreach Team 12 hours later. The patient died one day later. Respiratory rate 344 83.5 Answers may be multiple n=412; 57 not answered The Reviewers were of the opinion that had a diagnosis of sepsis been considered and documented the patient’s Investigations care might have been more comprehensive and quicker. Blood cultures were taken in 366/477 (77%) patients (Table 6.7). Of those patients who had blood cultures taken It should be expected that observations would be frequent the Reviewers considered there to be delays in 52/298 (17%) once a diagnosis of sepsis had been made, and in the vast patients (Table 6.8). Tissue cultures and fluid cultures were majority of cases, vital signs were monitored to some extent taken in 43/268 (16%) and 307/449 (68.3%) of patients (507/525; 97%; Table 6.5). respectively.

Table 6.5 Vital signs taken at time of sepsis Table 6.7 Blood cultures taken identification Blood cultures Number of % Vital signs Number of % patients patients Yes 366 76.7 Yes 507 96.6 No 111 23.3 No 18 3.4 Subtotal 477 Subtotal 525 Insufficient data 74 Insufficient data 26 Total 551 Total 551 Table 6.8 Delay in blood cultures being taken – However, review of the detail of observations taken showed Reviewers’ opinion that the patients in this study appeared not to have had Delay in blood cultures Number of % adequate observations made (Table 6.6). The Glasgow patients Coma Score and AVPU scale are simple means of assessing Yes 52 17.4 a patient’s conscious level. A reduction in conscious level No 246 82.6 and state of mind is a sign of worsening sepsis. Only 50% Subtotal 298 of patients had their conscious level assessed at the time of Insufficient data 68 diagnosis. Equally concerning is that 1 in 4 patients did not Total 366 have their respiratory rate recorded.

72 6

Number of patients

12

10

8

6

4

2

0 30 mins- >1-2 hours >2-3 hours >3-6 hours >6-12 hours >12-24 hours >24 hours 1 hour

Figure 6.5 Time delay from sepsis diagnosis to blood cultures being taken – Clinician questionnaire

The time duration from sepsis diagnosis to blood cultures At the time sepsis was diagnosed, blood gases were taken for those cases where the Reviewers considered that there from 375/509 (74%) patients in the study (Table 6.10). Table was a delay in them being taken is shown in Figure 6.5. 6.11 shows the time frame when blood gases were taken. Most of these patients 36/39 had blood cultures taken Table 6.10 Blood gases taken outside the one hour standard set for administration of antimicrobials in patients with sepsis.1 Blood cultures are Blood gases Number of % more likely to identify a pathogen if they are taken before patients the administration of antimicrobials. Yes 375 73.7 No 134 26.3 Of the 111 patients who did not have blood cultures Subtotal 509 taken, 48 had fluid cultures and 43 had tissue cultures. Insufficient data 42 The Reviewers noted three patients who did not have any Total 551 cultures taken and there was insufficient data to comment for the remaining 20. Table 6.11 Time frame for taking blood gases

Table 6.9 Tissue cultures taken Hours Number of % patients Tissue cultures Number of % patients Immediately 132 52.0 Yes 43 16.0 within 1 hour 69 27.2 No 225 84.0 >1-4 hours 24 9.4 Subtotal 268 >4-8 hours 8 3.1 Insufficient data 283 >8-12 hours 10 3.9 >12-24 hours 6 2.4 Total 551 >24 hours 5 2.0 Subtotal 254 Not answered 121 Total 375

73 First identification of sepsis

Table 6.12 shows investigations carried out at the time Table 6.13 Investigations missed – Reviewers’ sepsis was diagnosed according to the Reviewers. The opinion ‘Sepsis Six’ is a set of interventions that if delivered in the Investigations missed Number of % first hour of care can lead to improved outcome.23 Within patients these interventions the measurement of serum lactate Yes 198 39.1 and haemoglobin are included. The majority of patients No 308 60.9 had haemoglobin measured as part of a full blood count. Subtotal 506 However, only 322/522 (62%) had lactate recorded, despite Insufficient data 45 74% of patients having blood gases that would normally include lactate measurement. Total 551

Table 6.12 Investigations carried out Table 6.14 Investigations delayed – Reviewers’ opinion Investigations carried out Number of % patients Investigations delayed Number of % Full blood count 490 93.9 patients Urea and electrolytes 491 94.1 Yes 190 38.3 Liver function tests 388 74.3 No 306 61.7 Amylase 114 21.8 Subtotal 496 CRP 396 75.9 Insufficient data 55 Ultrasound 29 5.6 Total 551 Urine analysis 234 44.8 CT scan 101 19.3 Table 6.15 shows the detail of the 246 (44%) cases where Lactate 322 61.7 investigations were either missed or delayed in the opinion of the Reviewers. Estimated glomerular filtration rate 174 33.3

Chest X-ray 364 69.7 Table 6.15 Missed/delayed investigations Coagulation screening 230 44.1 Investigation Number of % Other 48 9.2 patients Answers may be multiple n=522 Imaging 42 17.1 Urine analysis 23 9.3 The Reviewers considered that there were investigations Blood cultures 27 11.0 missed in 198/506 (39%) patients (Table 6.13) and delayed Sputum culture and analysis 11 4.5 in 190/496 patients (38%) (Table 6.14). Appropriate Arterial blood gases 7 2.8 diagnosis and treatment are dependent on a full array of investigations. Wound swabs 6 2.4 Lactate 4 1.6 Lumbar puncture 3 1.2 Urinary output 2 <1 Physical examination of specific 1 <1 area Amylase 1 <1 Answers may be multiple n=246

74 6

By far the most commonly delayed or missed investigation Table 6.18 Investigations not done to identify was radiological imaging. In contrast with the Reviewers, source of sepsis that should have been the treating clinicians at the hospital considered that only Investigations omitted Number of % 12% (80/649) of patients had had an investigation delayed patients or omitted (Table 6.16). Yes 113 22.8 No 382 77.2 Table 6.16 Investigations omitted/delayed at the time sepsis was identified - opinion of the treating Subtotal 495 clinician Insufficient data 56 Omitted or delayed Number of % Total 551 investigations patients Yes 80 12.3 In 163/551 (29%) patients, the Reviewers believed that No 569 87.7 investigations carried out to identify the source of sepsis Subtotal 649 were either delayed or not done. The source of sepsis was Insufficient data 61 successfully identified in 434/493 (88%) patients (Table 6.19). Of these it was stated by the Reviewers that it was Total 710 identified within an appropriate time frame in 340/421 (81%) cases (Table 6.20). Reviewers were of the opinion that there were delays in the investigations carried out to identify the source of infection Table 6.19 Source of sepsis identified and sepsis in 101/505 (20%) patients and that there were Identified Number of % investigations to identify the source of infection omitted patients that should have been performed in 113/495 (23%) patients Yes 434 88.0 (Table 6.17, Table 6.18). Early source identification is No 59 12.0 important if sepsis is to be treated promptly. Subtotal 493 Insufficient data 58 Table 6.17 Delay in investigations to identify the source of sepsis – Reviewers’ opinion Total 551 Delay in investigations Number of % patients Table 6.20 Sepsis source identified within Yes 101 20.0 appropriate timeframe – Reviewers’ opinion No 404 80.0 Timely identification Number of % patients Subtotal 505 Yes 340 80.8 Not applicable 5 No 81 19.2 Insufficient data 41 Subtotal 421 Total 551 Not applicable 1 Insufficient data 12 Total 434

75 First identification of sepsis

C A S E S T U D Y 7 Key Findings

An elderly patient was admitted to hospital following a minor trauma and developed a large retroperitoneal • There was a delay in identifying sepsis in 182/505 (36%) haematoma. The patient was on warfarin for atrial of cases, severe sepsis in 167/324 (51.5%) cases and fibrillation and it was subsequently found that their septic shock in 63/193 (32.6%) cases according to the INR was >8. Whilst in hospital the patient developed Reviewers a hospital-acquired pneumonia. The patient became • According to the Reviewers 128/479 (26.7%) patients hypotensive and oliguric and it was 24 hours before had an EWS screening tool used to aid the diagnosis of any recognition of their condition was acknowledged. sepsis Documentation was poor and sepsis was not • The use of a screening tool/ EWS was associated with mentioned in the case notes despite clear evidence fewer delays in identifying severe sepsis (55% without from physiological observations and blood results that vs. 35% with) this patient had severe sepsis. Despite the delay in recognition and treatment the patient was discharged • Only 52.9% (218/412) patients had their GCS/AVPU home from hospital, but with significant cognitive assessed at the time of diagnosis impairment. • Only 322/522 (61.7%) patients had a record in the notes that lactate had been measured The Reviewers commented on the delay in recognition, • Investigations considered essential in the diagnosis of poor documentation and failure to mention the word sepsis were missed in 198/506 (39.1%) patients and ‘sepsis’. They considered that the clinical care could delayed in 190/496 (38.3%) have been greatly improved.

76 7

Initial management of sepsis Back to contents

The clinical teams overseeing management of the patients Table 7.2 Timely escalation/commencement of with sepsis are listed in Table 7.1. As was to be expected treatment – Reviewers’ opinion acute medical teams predominantly managed the patients Timely escalation/ Number of % with sepsis in this study. commencement of treatment patients Yes 406 80.6 Table 7.1 Specialty of team overseeing management No 98 19.4 of the patient following diagnosis Subtotal 504 Specialty Number of % Insufficient data 47 patients Acute/general medicine 270 50.1 Total 551 Other specialist medical team 43 8.0 General surgery 88 16.3 Table 7.3 If not timely – patient deteriorated whilst Other specialist surgery 37 6.9 waiting for treatment Emergency medicine 57 10.6 Patient deteriorated Number of % patients Critical care outreach 76 14.1 Yes 51 59.3 Critical care - Level 2 69 12.8 No 35 40.7 Critical care - Level 3 79 14.7 Subtotal 86 Other 33 6.1 Insufficient data 12 Answers may be multiple n=539 Total 98

The Reviewers considered that 406/504 (80.6%) patients had Table 7.4 Outcome affected by a delay in escalation/ timely escalation of care following diagnosis (Table 7.2). In treatment – Reviewers’ opinion those for whom the escalation was not timely the Reviewers Outcome affected Total considered the patient to have deteriorated in 51/86 cases (Table 7.3) and that the final outcome was affected in Yes 20 20/39 patients (Table 7.4); 12/20 of these patients died No 19 prior to discharge. Although there is insufficient evidence Subtotal 39 presented here to determine a causal link between a lack of Insufficient data 12 timely escalation and deterioration, the message that timely Total 51 treatment of sepsis is paramount should be reinforced.

77 Initial management of sepsis

Fluid resuscitation is another pillar of the ‘Sepsis Six’ Fluid balance charts aid the appropriate resuscitation approach to managing sepsis. The Reviewers identified of patients with sepsis. However the Reviewers were of that 9 out of 10 patients required fluids. Of those requiring the opinion that there was room for improvement in the fluids 82.8% (341/412) received fluids promptly, in patients’ fluid management in 203/447 (45.4%) cases 11.9% (49/412) administration was delayed and 5.3% (Table 7.7) and in half the patients, where there was room (22/412) received no intravenous fluids. Contemporary for improvement, the reason for improvement was poor understanding of the pathophysiology of sepsis supports documentation of fluid balance (Table 7.8). intensive fluid resuscitation in the initial phase. SIRS and sepsis incite widespread inflammatory responses at Table 7.7 Room for improvement in fluid tissue and cellular levels altering homeostasis. Resultant management – Reviewers’ opinion circulatory abnormalities lead to an imbalance between Room for improvement Number of % oxygen delivery and demand, worsening end organ injury patients and failure. Although adequate fluid resuscitation makes Yes 203 45.4 physiological sense, the optimal amount, and type of fluid No 244 54.6 remain unclear.46 Whichever fluid is chosen, resuscitation Subtotal 447 should combine clinical assessment, such as signs of tissue Insufficient data 104 perfusion with haemodynamic monitoring. Total 551 Table 7.5 IV fluid resuscitation required Table 7.8 Principal reasons for room for Fluid required Number of % improvement in fluid management patients Fluid resuscitation required 453 89.9 Reason Number of % patients Fluid resuscitation not required 51 10.1 Documentation of fluid balance 95 51.6 Subtotal 504 Delay in commencing fluid 71 38.6 Insufficient data 47 resuscitation Total 551 Monitoring-frequency/type 58 31.5 Type of fluid 20 10.9 Table 7.6 Delivery of the fluids administered Too slow rate of IV fluids 13 7.1 Delivery Number of % Delay commencing vasopressors 11 6.0 patients Overloading with fluids 5 2.7 Promptly received 341 82.8 Catheterisation 5 2.7 Received but delayed 49 11.9 Other documentation 5 2.7 Not received 22 5.3 Blood products 3 1.6 Subtotal 412 Lack of planning for fluid 3 1.6 Insufficient data 41 management Total 453 Other 41 22.3 Answers may be multiple n=184; 19 not answered

78 7

The Reviewers considered that in those cases where there was Clinicians who cared for the patients identified the top ten room for improvement in fluid balance, the final outcome sources of infection as seen in Table 7.12. Respiratory and could have been affected in 1 in 5 patients (Table 7.9). urinary tract were at the top of the possible causes, which is in line with published data. Table 7.9 Patient’s outcome have been affected by poor fluid balance – Reviewers’ opinion Table 7.12 Presumed source of infection – Clinician Outcome affected Number of % questionnaire patients Infection Number of % Yes 29 20.7 patients No 111 79.3 Respiratory tract 297 42.8 Subtotal 140 Urinary tract 168 24.2 Insufficient data 52 Acute abdominal/upper 127 18.3 gastrointestinal tract Total 192 Skin/soft tissue 65 9.4 Post operative 43 6.2 The administration of oxygen is another of the interventions Intracranial/ear, nose and throat 21 3.0 of the ‘Sepsis Six’. Despite a diagnosis of sepsis the Perianal/ischio-rectal/ lower 17 2.4 Reviewers considered that 20% (100/502) of patients did gastrointestinal tract not require oxygen (Table 7.10). Bone/joint 11 1.6 Table 7.10 Oxygen requirement Endocarditis 9 1.3 Oxygen required Number of % Implantable device 8 1.2 patients Gynaecological/sexually 7 1.0 Oxygen therapy required 402 80.1 transmitted infection Oxygen therapy not required 100 19.9 Other 41 5.9 Subtotal 502 Answers may be multiple n=694; not answered in 16 Insufficient data 49 In this study a pathogen was identified in 198/481 (41%) Total 551 patients (Table 7.13) and the commonest pathogens are shown in Figure 7.1 overleaf. Of those who required oxygen 334/371 (90%) received it promptly, in 24/371 (6.5%) administration was delayed and Table 7.13 Pathogen identified in 13/371 (4%) no oxygen was administered (Table 7.11). Pathogen Number of % patients Table 7.11 Oxygen delivery when required – Yes 198 41.2 Reviewers’ opinion No 283 58.8 Delivery Number of % Subtotal 481 patients Insufficient data 70 Promptly received 334 90.0 Total 551 Delayed 24 6.5 Not received 13 3.5 Subtotal 371 Insufficient data 31 Total 402

79 Initial management of sepsis

Percentage of patients

30 51 25

39 20 29 15

10 14 10 9 5 9 6 4 3 3 2 2 0

Proteus Candida Klebsiella Legionella C. Difficile Morganella Streptococcus PseudomonasEnterococcus Haemophilus Staphylococcus E.Coli & coliforms Other gram negative Figure 7.1 Top 13 pathogens identified in patients in the study

Escherichia coli and other coliform bacteria were the most Table 7.14 Most common samples used to isolate a commonly isolated pathogen followed by streptococcus specimen and staphylococcus. Fungi were rare in this cohort with Sample Number of % candida isolated from a minority of patients. The causative patients organisms for sepsis have evolved over many years. The Blood cultures 87 47.3 original studies of sepsis bore out that Gram-negative Urine 49 26.6 bacteria were among the most common causes of sepsis. It Sputum 21 11.4 is now recognised that sepsis may occur from any bacteria, Wound swab 10 5.4 as well as from fungal and viral organisms. More recent Answers may be multiple n=184; 8 not answered epidemiology studies reveal that Gram-positive bacteria have become the most common cause of sepsis in the past 25 years. According to the most recent estimates in Despite 95% (323/339) of hospitals including the sepsis, there are approximately 200,000 cases of Gram- administration of IV antimicrobials in their sepsis protocols positive sepsis each year, compared with approximately and 95% (305/321) of those protocols specifying that 150,000 cases of Gram-negative sepsis.47 While bacterial antimicrobials should be given within one hour, only causes of sepsis have increased with the general increases 226/361 (63%) patients received antimicrobials within one in incidence, fungal causes of sepsis have grown at an even hour (Table 7.15). more rapid pace. This may represent a general increase in hospital-acquired cases of sepsis. Kumar et al. brought the importance of early antimicrobial therapy in patients with septic shock to the forefront in Where specimens were successfully isolated, it was most 2006. Kumar and his colleagues completed a retrospective commonly from blood (47%) urine (27%) or sputum (11%) cohort study of 2,731 adult patients with septic shock, (Table 7.14). This is in line with the presumed sources of examining mortality in patients who received antimicrobials infection listed in Table 7.12. after the onset of recurrent or persistent hypotension. They found that administration of an appropriate antimicrobial within one hour of identified hypotension resulted in a survival rate of 79.9%. Each hours’ delay in administration of an antimicrobial resulted in a 7.6% decrease in survival.48

80 7

Number of patients No sepsis on admission Sepsis Severe sepsis Septic shock 50

45

40

35

30

25

20

15

10

5

0 <30 mins >30 mins - >1 hour - >2 hour - >6 hour - >12 hours <1 hour <2 hours <6 hours <12 hours

Time from sepsis diagnosis to first dose of antimicrobials Figure 7.2 Time delay in administration of antimicrobial and sepsis status on admission to hospital (Reviewer’s opinion)

Table 7.15 Time delay for administration of Similarly, in 2010, the International Surviving Sepsis antimicrobials Campaign (SSC) published results in over 15,000 episodes Time first dose of Number of % showing that delivery of early antimicrobials (at that stage antimicrobials given from first patients within 3 hours) was independently associated with survival diagnosis of sepsis in a risk-adjusted model (odds ratio 0.86), but was achieved <30 minutes 118 32.7 in only 67% of cases.11 Figure 7.2 shows the time delay >30 mins - < 1 hour 108 29.9 in administering antimicrobials to patients, grouped by >1 hour - < 2 hours 70 19.4 severity of sepsis. It can be seen that whilst the majority of >2 hours - < 6 hours 43 11.9 patients in all groups received antimicrobials within 6 hours >6 hours - < 12 hours 17 4.7 of diagnosis, there were 21 patients with septic shock on > 12 hours 5 1.4 admission who received antimicrobials more than 1 hour Subtotal 361 after diagnosis and 13 patients with severe sepsis or septic Time antimicrobials administered 100 shock on admission who had a delay of more than 6 hours not documented in the delivery of antimicrobials. Time sepsis diagnosed not 32 documented Neither time documented 7 Not applicable - already on 33 antimicrobials Not applicable - antimicrobials 1 not given Insufficient data 17 Total 551

81 Initial management of sepsis

Table 7.16 Length of time from first recorded Table 7.17 Avoidable delay in administering arrival (triage) in the emergency department to the antimicrobial – Reviewers’ opinion first dose of antimicrobial in patients considered Avoidable delay Number of % by the Reviewers to have sepsis on arrival in the patients emergency department. Yes 114 29.2 Time from first recorded Number of % arrival in the emergency patients No 277 70.8 department to first dose of Subtotal 391 antimicrobial Insufficient data 160 Already received first dose of 7 2.8 Total 551 antimicrobial on arrival 0-30 minutes 35 13.8 >30minutes- 1 hour 29 11.5 Table 7.18 Reason for delay in administering >1 hour- 3 hours 89 35.2 antimicrobials >3 hours to 6 hours 42 16.6 Reason Number of % >6 hours- 12 hours 32 12.6 patients >12- 24 hours 7 2.8 Delay in prescribing 40 39.2 >24 hours 12 4.7 Lack of escalation 17 16.7 Subtotal 253 Unclear from the notes 14 13.7 Data missing/no sepsis in ED 25 Awaiting source confirmation 10 9.8 Total 278 Communication between teams 9 8.8 e.g. handover Table 7.16 shows the time delay from first arrival in the Delay in sepsis recognition 9 8.8 emergency department to the first dose of antimicrobial Delay review by senior staff 8 7.8 being administered in patients who the Reviewers felt had Delay in administration after 6 5.9 evidence of sepsis on arrival to the emergency department. prescribing One fifth of this group (51/253; 20%) had a delay of more Not prescribed a stat dose 6 5.9 than 6 hours in the administering of antimicrobials. This is a Awaiting microbiology review 3 2.9 combination of the delay in identification of sepsis and the IV access problems 2 2.0 delay in administering antimicrobials once the diagnosis has Other clinical reason 2 2.0 been made. Pharmacy delay 0 0.0 The Reviewers considered that there was an avoidable delay Answers may be multiple n=102; not answered in 12 in the administration of antimicrobials in 114/391 (29%) patients. The principle identifiable cause of delay was in The Reviewers were of the opinion that a delay in the prescribing (Table 7.17) and Table 7.18 details the reason administration of antimicrobials affected the outcome in given for the delay. nearly half (43/97; 44%) of those patients who did not receive antimicrobials in a timely manner (Table 7.19). There may be benefit from the extension of prescribing responsibilities to non-medically qualified healthcare In those in whom the Reviewers considered that a delay may professionals. Triage nurses and nurse practitioners in AMUs have affected outcome the Reviewers’ reasons fell into the would be ideally placed to ensure patients with sepsis broad themes of wrong route of administration and wrong received prompt antibiotic therapy. antimicrobial given later than considered appropriate.

82 7

Table 7.19 Outcome affected by a delay in Table 7.21 Patient started on a sepsis care bundle antimicrobials – Reviewers’ opinion following diagnosis – Clinician questionnaire Outcome affected Number of % Sepsis care bundle Number of % patients patients Yes 43 44.3 Yes 207 39.4 No 54 55.7 No 318 60.6 Subtotal 97 Subtotal 525 Insufficient data 17 Not answered 185 Total 114 Total 710

Table 7.20: Patient started on a sepsis care bundle following diagnosis – Reviewers’ opinion hospitals claiming to have specific sepsis care bundles (Chapter 2). Sepsis care bundle Number of % patients The clinicians completing the questionnaire similarly Yes 135 31.1 reported that 207/525 (39%) patients were managed on a No 299 68.9 care bundle following diagnosis (Table 7.21). Subtotal 434 Insufficient data 117 Figure 7.3 demonstrates the positive effect of care bundles Total 551 on timely treatment, antimicrobial administration, timely escalation, prompt fluid administration, prompt oxygen Only one in three patients (135/434; 31%) were started on administration, documentation, blood cultures and gases a sepsis care bundle (Table 7.20), despite 358/544 (66%) being taken, and the identification of a source of sepsis.

No care bundle Care bundle

79.5% 23.0% 33.1% 38.1% 12.5% 60.0% 26.3% 13.2% 15.3% 28.1% 19.2% 18.5% 6.3% 9.9% 5.3% 10.2%

n=92 n=116 n=299 Blood cultures Less than good Delay in escalation Fluids receiveddelayed/not n=61 received n=37 Oxygen delayed/not not taken n=84 antimicrobials n=94 Blood gases not taken Delay in administration of Identification of source of infection delayed n=41 documentation of sepsis Figure 7.3 Use of care bundles

83 Initial management of sepsis

With significant advancement in the care of patients with Table 7.22 How the antimicrobial was chosen sepsis and implementation of guidelines for bundled care, Antimicrobial choice Number of % investigators involved in the Surviving Sepsis Campaign patients demonstrated a decline in 28 day mortality from 37% According to local hospital policy 191 36.0 to 30.8% over two years. Compliance with the initial six Previous culture results 25 4.7 hour bundle, termed the resuscitation bundle, increased Based on site of infection 121 22.8 from 10.9% to 31.3% of subjects over the two years after initiation of the SSC guidelines. It is difficult to know Administered broad spectrum 128 24.1 antibiotics if the improved outcomes that were observed should be attributed to the SSC guidelines, or to a heightened Rationale not documented 181 34.1 appreciation by medical teams that the treatment of sepsis Other 33 6.2 is time sensitive.1 More recently, the National Emergency Answers may be multiple n=531, not answered in 20 Laparotomy Study49 demonstrated the relationship between time from onset of hypotension, time to administration of Table 7.23: Appropriateness of the choice of antimicrobials and survival fraction. antimicrobial therapy – Reviewers’ opinion Appropriate antimicrobial Number of % Antimicrobial stewardship patients Yes 472 90.6 The term ‘antimicrobial stewardship’ is defined as ‘an No 49 9.4 organisational and by implication, healthcare-system-wide Subtotal 521 approach, to promoting and monitoring the correct use Insufficient data 30 of antimicrobials to preserve their future effectiveness’ and Total 551 the recently published NICE guideline recommends specific approaches to the prescription of antimicrobials which should be adhered to.12 Table 7.24 Correct dose of antimicrobial therapy – Reviewers’ opinion The rationale behind the choice of antimicrobial in this study Correct dose of antimicrobial Number of % is outlined in Table 7.22. In 1 in 3 patients (181/531; 34%) patients no rationale for prescription was documented. However, Yes 405 97.8 the Reviewers considered that only 1 in 10 (49/521; 9%) No 9 2.2 patients had been started on an inappropriate antimicrobial Subtotal 414 (Table 7.23). The correct dose was prescribed in 98% Insufficient data 58 (405/414) of patients (Table 7.24). Total 472 Much research from the last decade has highlighted the strong relationship between the choice of empiric In this study, a consultant microbiologist was consulted on antimicrobial therapy and the risk of death among the suitability of therapy only in half the patients (244/471; patients hospitalised with serious infections. Most studies 52%; Table 7.25). However, 317/404 (78.5%) patients had suggest that the risk of hospital death in association with a regular review of their antimicrobial therapy (Table 7.26), inappropriate initial antibiotic therapy goes up twofold which was most commonly undertaken by a microbiologist to fourfold when compared with patients who receive (153/276; 55%) (Table 7.27). appropriate coverage.50-55

84 7

Table 7.25 Consultation with a microbiologist Table 7.27 Clinical specialty that conducted the review of antimicrobial therapy Consultation with a Number of % microbiologist patients Specialty Number of % Yes 244 51.8 patients No 227 48.2 Microbiologist 153 55.4 Subtotal 471 Critical care 34 12.3 Insufficient data 80 Medical team 83 30.1 Total 551 Surgical team 10 3.6 Pharmacy 4 1.4 Other 4 1.4 Table 7.26 Regular review of antimicrobial therapy Subtotal 276 Regular review Number of % Insufficient data 16 patients Total 317 Yes 317 78.5 Answers may be multiple No 87 21.5 Subtotal 404 Escalation of antimicrobials was considered in 85% Insufficient data 147 (358/420) of patients. De-escalation was considered in Total 551 74% (289/389) and the duration of therapy considered in 80% (329/413) of patients (Table 7.28 and Figure 7.4).

Table 7.28 Details of the antimicrobial therapy considered Consideration given to Yes %Yes No %No Subtotal Insufficient Total data 1) Escalation of antimicrobial therapy 358 85.2 62 14.8 420 131 551 2) De-escalation of antimicrobial therapy 289 74.3 100 25.7 389 162 551 3) Duration of antimicrobial therapy 329 79.7 84 20.3 413 138 551

Percentage Cases consideration given Cases consideration not given

90

80

70

60

50

40

30

20

10

0 1) Escalation of antimicrobial 2) De-escalation of antimicrobial 3) Duration of antimicrobial therapy (n=420) therapy (n=389) therapy (n=413)

Figure 7.4 Details of antimicrobial therapy considered

85 Initial management of sepsis

Following microbiology review, changes were made in Once identified, the focus of infection should be drained 360/482 (74%) of this group (Table 7.29). or removed with the least physiologic insult. The recent National Emergency Laparotomy Audit demonstrated Table 7.29 Antimicrobial therapy modified following that both early source control and early antimicrobial microbiology review administration dramatically improves survival.49 Therapy modified Number of % patients In this study a source of infection was identified in 434/493 Yes 360 73.6 (88 %) cases (Table 7.30). In 16 of 49 cases the Reviewers No 129 26.4 considered more could have been done to identify the Subtotal 489 source (Table 7.31). Only in 29% of cases was the source amenable to an immediate procedure for control (Table 7.32). Insufficient data 62

Total 551 Table 7.30 Source of infection was identified Source identified Number of % Establishing an optimal antibiotic dosage regimen is patients important, but so is the presence of an effective de- Yes 434 88.0 escalation strategy to shorten unnecessary antibiotic No 59 12.0 exposure by streamlining therapy and narrowing antibiotic Subtotal 493 choices based on a patient’s culture result. After the Insufficient data 58 initiation of broad-spectrum antibiotics, de-escalation can be typically conducted in several ways: the antibiotics can be Total 551 streamlined to more narrow-spectrum agents once culture and susceptibilities are available, the dosage if initially high Table 7.31 If ‘No’ could more have been done to can be de-escalated to a standard dosage for a susceptible identify the source organism, or antibiotics can be discontinued altogether if More could have been done Number of subsequent data reveal it is unlikely that the patient had patients an infection in the first place.56 This study demonstrates Yes 16 that clinicians have heeded the need for antimicrobial No 30 stewardship but there is ongoing room for improvement. Subtotal 46 Insufficient data 13 Source of infection and its control Total 59 Infection source control dates back to the origins of medicine.57 The need to drain abscesses and remove foreign Table 7.32 Infection source was amenable to bodies has been recognised since the fourth century BC, immediate control and the modern management of sepsis still depends on Amenable to control Number of % such surgical therapy. Identifying a source of infection patients and removing it, if possible, remains a core principle. The Yes 137 28.7 Surviving Sepsis Campaign guidelines recommend that No 341 71.3 the source of infection be confirmed within 6 hours of Subtotal 478 presentation.8 This requires blood cultures to be taken prior Not applicable 17 to the administration of antimicrobials and often requires Insufficient data 56 ultrasound or CT imaging; other cultures (e.g. urine, cerebrospinal fluid, synovial fluid) may also be required. Total 551

86 7

The procedures performed to control the source of infection Table 7.35 Reasons for delay in controlling the are shown in Table 7.33. In those patients in whom a source source of infection – Reviewers’ opinion was amenable to control the Reviewers were of the opinion Reason for delay Number of that control was delayed in 43% (Table 7.34), the reasons patients for delay are listed in Table 7.35. ‘Other’ reasons for delay Patient too unwell to tolerate surgery 7 included the patient initially being treated conservatively, Lack of beds 2 missed diagnosis, delayed senior review, patient refusal and Out of hours/weekend 4 equipment failure (Table 7.36). Delay in investigations 15 Lack of available staff 7 Table 7.33 Procedure performed to control the source of infection Delay in identifying source 16 Procedure performed Number of % Lack of specialist 5 patients Patient reasons - refusal, consent 3 Laparotomy +/- wash out 35 25.5 Next scheduled list 2 Abscess drainage under 8 5.8 Reason not documented 4 interventional radiology Answers may be multiple n=137 Chest drain 7 5.1 The Reviewers were of the opinion that delay in source Nephrostomy 7 5.1 control affected the outcome in 33/47patients (Table 7.36). Catheter irrigation/replacement 6 4.4 Laparosopy and wash out 5 3.6 Table 7.36 Delay in source control affected the Endoscopic retrograde 4 2.9 outcome – Reviewers’ opinion cholangiopancreatography (ERCP) Outcome affected Number of Line/peg replacement 4 2.9 patients Amputation 3 2.2 Yes 33 Joint debridement/washout 3 2.2 No 14 Gallbladder drainage 3 2.2 Subtotal 47 Ventricular drain 2 1.5 Insufficient data 8 Other 8 5.8 Total 55 Answers may be multiple n=137

C A S E S T U D Y 8 Table 7.34 Delay in source control – Reviewers’ opinion An elderly patient with prostate cancer was admitted Delay Number of % with a urinary tract infection and signs of sepsis. The patients diagnosis of sepsis due to perinephric abcess was Yes 55 42.6 made and within one hour the patient had undergone No 74 57.4 appropriate imaging and 90 minutes after the imaging Subtotal 129 the patient underwent a percutaneous nephrostomy. Insufficient data 8 The patient was discharged from hospital 9 days later. Total 137 The Reviewers considered that this demonstrated the value of early source control in sepsis.

87 Initial management of sepsis

Patient information Table 7.37 Evidence of discussion between healthcare professionals and relatives/carer/patient The patient notes contained evidence that discussions had Evidence Number of % taken place between healthcare providers and the patient patients and/or their relatives in two thirds (299/452; 66%) of cases Yes 299 66.2 (Table 7.37). However 3 in 10 patients appeared not to have No 153 33.8 had any discussion about the condition. Where discussions Subtotal 452 took place it appears that the discussions were more often Insufficient data 92 with patients’ relatives than with the patients (Table 7.38). Not applicable 7 Total 551

Table 7.38 Detail of discussions with relatives/carer/patient Evidence in the notes Patient Relative Yes Yes% No Subtotal Yes Yes% No Subtotal 1) Diagnosis of sepsis: 94 49.7 95 189 165 67.9 78 243 2) Cause of sepsis 108 56.8 82 190 180 72.9 67 247 3) Regularly updated treatment plan 100 59.2 69 169 159 75.0 53 212 4) Possible outcome for the patient 89 53.3 78 167 187 82.4 40 227 5) Rehabilitation plan 55 40.4 81 136 53 36.6 92 145

Management on general wards Thirteen patients received inotropes in the general ward environment of which evidence of adequate monitoring Just under half of the patients in this study were managed could be found in only five. The Reviewers considered on a general ward throughout their episode of sepsis (Table the administration of inotropes on a general ward was a 7.39). This group was managed with input from Critical holding manoeuvre prior to transfer to a high care area Care Outreach Teams, critical care and microbiology (data in five of the 13 patients. The Reviewers considered that not shown). all 13 patients had received inotropes in a timely manner. However, they considered that the administration of Table 7.39 Entire episode managed on a general inotropes on a general ward was only appropriate in ward 5/13 patients (the same five who were given inotropes Episode managed on general Number of % as holding measure). ward patients Yes 257 47.2 No 288 52.8 Subtotal 545 Insufficient data 6 Total 551

88 7

Table 7.40 Managing the entire patient’s episode on The Reviewers were of the opinion that 93% (209/224) of the ward was appropriate – Reviewers’ opinion patients who were managed on the general ward were Appropriate management Number of % in the correct location (Table 7.40). The reasons as to patients why 6.7% were not in the correct location can be broadly Yes 209 93.3 grouped into: delay in investigations, inadequate monitoring No 15 6.7 and delays in delivering care (data not shown). Subtotal 224 Table 7.41 summarises the areas where the Reviewers Insufficient data 33 considered improvements could have been made in the Total 257 initial management of patients and whether they thought the patient’s outcome had been affected. When ranked in order of influence on outcome, failure to deal with the source of infection ranked highest (87%) with inadequacies in review and delay in diagnosis the next two most influential areas for improvement.

Table 7.41 Reasons for room for improvement in the initial management of patients – Reviewers’ opinion Listed reason for room for If YES - outcome affected improvement Yes Yes No Subtotal ID Outcome Outcome No Subtotal ID Total % affected affected % Failure to adhere to 201 76.4 62 263 29 116 67.1 57 173 28 292 sepsis 6 pathway Documentation 183 69.6 80 263 29 58 39.7 88 146 37 292 Delay in diagnosis 173 65.8 90 263 29 107 66.9 53 160 13 292 of sepsis Communication 137 53.9 117 254 38 25 21.0 94 119 18 292 with patient/ relatives Inadequacies in 129 51.4 122 251 41 84 73.7 30 114 15 292 review Inadequacies in 99 39.6 151 250 42 99 39.6 151 250 15 292 monitoring Delay in diagnosis 92 35.9 164 256 36 59 71.1 24 83 9 292 of infection Failure to deal with 78 31.7 168 246 46 55 87.3 8 63 15 292 source of infection within acceptable timeframe Other 51 64.6 28 79 213 19 70.4 8 27 24 292 ID = insufficient data

89 Initial management of sepsis

Escalation and critical care management Table 7.44 Timely referral to the Critical Care Outreach Team Critical Care Outreach Teams, Rapid Response Teams or Timely referral to CCOT Number of % Medical Emergency Teams, depending on the geographical patients location, have become increasingly involved in sharing their Yes 287 83.9 expertise of critical care by reviewing and treating patients No 55 16.1 early on in their acute illness, on the ward as well as in the Subtotal 342 critical care unit, in order to prevent further deterioration Not applicable 10 58 and death. The benefit of Critical Care Outreach Teams Insufficient data 54 has been demonstrated to reduce hospital morbidity and Total 406 mortality.59 In this cohort of patients 79% (406/513) of patients were referred to Critical Care Outreach Teams. Table 7.45 Reasons why the referral to the Critical In the group that was not referred 48/107 patients were Care Outreach Team was not timely – Reviewers’ admitted directly to Level 3 care (Table 7.42). opinion Reason why referral was not timely Number of Table 7.42 Patient referred to the Critical Care patients Outreach Team Inaccurate calculation of early warning 1 score Referred to CCOT Number of % Did not respond to high early warning 12 patients score/did not appreciate severity of illness Yes 406 79.1 Insufficient seniority of review 12 No 107 20.9 Insufficient frequency of clinical review 5 Subtotal 513 Insufficient monitoring of observations 3 Not applicable 9 Critical Care Outreach Team not available 1 out of hours Insufficient data 29 Other 4 Total 551 Insufficient data 19 Answers may be multiple n=55

Table 7.43 Trigger for the referral to the Critical Care Table 7.46 Critical Care Outreach Team arrived Outreach Team promptly following contact with them – Reviewers’ Trigger for referral Number of % opinion patients Prompt arrival Number of % Automatic call to the Critical Care 62 18.6 patients Outreach Team – early warning Yes 237 88.8 score trigger No 30 11.2 Direct referral by clinicians on 128 38.3 Subtotal 267 ward Not applicable 13 Direct referral by nursing staff 33 9.9 Insufficient data 126 Peri/cardiac arrest 12 3.6 Total 406 Manual early warning score 69 20.7 calculation The commonest trigger for a referral to the Critical Care Other 30 9.0 Outreach Team was direct clinician referral (Table 7.43). The referrals were considered to be timely in 84% (287/342) of cases Subtotal 334 (Table 7.44) and if not timely this was due most commonly to Insufficient data 72 lack of senior review (Table 7.45). The Critical Care Outreach Total 406 Team arrived promptly on 89% of occasions (Table 7.46).

90 7

One in four patients had input from the Critical Care Table 7.50 Outcome of critical care referral Outreach Team before formal referral to critical care Outcome of referral Number of % (Table 7.47). patients Continued management on 28 10.1 Table 7.47 Input from the Critical Care Outreach general ward - no change Team prior to referral to critical care Transferred to Level 3 care 103 37.1 CCOT input Number of % Transferred to Level 2 care 102 36.7 patients Continued management on 34 12.2 Yes 64 22.5 general ward - advice from No 220 77.5 critical care Subtotal 284 Other 11 4.0 Not applicable 13 Total 278 Insufficient data 109 Table 7.51 Correct decision to admit the patient to Total 406 Level 3 care – Reviewers’ opinion Correct decision not to admit to Number of Sixty percent of the patients in this cohort were referred to critical care patients critical care (Table 7.48). Critical care services responded in Yes 47 an appropriate time frame in 93% of referrals (Table 7.49). No 10 Three quarters of the critical care referrals of patients with Subtotal 57 sepsis were then transferred to critical care (Table 7.50). Not answered 16 Table 7.48 Patient was referred to critical care Total 73 Referred to critical care Number of % patients The Reviewers considered that in 10 of 57 patients in whom Yes 278 58.8 information was available, that the decision not to admit to No 195 41.2 Level 3 was incorrect (Table 7.51). Subtotal 473 The reasons given included lack of available critical care beds Not applicable 10 dictating management decisions and missed opportunities Insufficient data 68 for intervention. Of those patients admitted to critical care Total 551 70% required support of their cardiovascular system, 78% of their respiratory system and 26% support of their renal Table 7.49 Timely response from critical care system (Table 7.52). Timely response Number of % Table 7.52 Systems requiring treatment in critical patients care Yes 222 93.3 System Number of % No 16 6.7 patients Subtotal 238 Cardiovascular system 119 70.0 Insufficient data 40 Respiratory system 132 77.6 Total 278 Renal System 44 25.9 Answers may be multiple, n=170, not answered in 35

91 Initial management of sepsis

Table 7.53 System, treatment given, timeliness System requiring treatment Cardiovascular system Respiratory system Renal system n=119 n=132 n=44 Treatment given Inotropes Ventilation Filtration Yes 93 89 29 % yes 94.9 90.8 76.3 No 5 9 9 Subtotal 98 98 38 Not applicable 0 3 2 Insufficient data 21 31 4 Timely 80 83 26 % timely 96.4 98.8 96.3 Not timely 3 1 1 Subtotal 83 84 17 Insufficient data 10 5 2

The Reviewers considered that in more than 90% of cases The Reviewers were of the opinion that monitoring was not the treatment was delivered in a timely fashion (Table 7.53). adequate in seven patients in whom the Reviewers felt that direct arterial blood pressure and cardiac output monitoring Table 7.54 Monitoring employed on the critical might have been beneficial (Tables 7.54 and 7.55). care unit

Monitoring employed Number of % C A S E S T U D Y 9 patients Arterial line 172 94.5 An elderly patient with a history of ischaemic heart Central venous catheter 123 67.6 disease, hypertension and 40 years of smoking was Central venous pressure 79 43.4 admitted with pneumonia and acute kidney injury. A measurement diagnosis of pneumonia and sepsis was made in the Cardiac output monitoring 21 11.5 emergency department. The patient was put on a Other 12 6.6 sepsis pathway and transferred to critical care. Within 30 minutes of arriving in hospital the ‘sepsis six’ had Answers may be multiple n=182; not answered in 23 been completed. Relatives were informed of the Table 7.55 Adequate monitoring in critical care – patient’s condition and escalation of care discussed. Reviewers’ opinion The patient required ventilatory support for three days in critical care. The patient made a full recovery and was Adequacy monitoring Number of % patients discharged from hospital 10 days later. Yes 176 96.2 The Reviewers considered that this patient had received No 7 3.8 prompt care that was at a standard that should be Subtotal 183 expected for all patients. The relatives were kept Insufficient data 22 informed throughout the admission and the severity of Total 205 the sepsis was identified early and documented clearly in the case notes.

92 7

There was no evidence of a regular review by a Table 7.58 Areas in need of improvement in critical microbiologist in 1 in 3 patients whilst in critical care (Table care – Reviewers’ opinion 7.56), which is comparable with the finding from the Areas for improvement Number of organisational questionnaire that there was incomplete patients coverage of microbiology input in critical care units. Documentation 19 Communication with patient/ relatives 10 Table 7.56 Regular review by microbiologist Treatment 9 Regular review Number of % Other 9 patients Communication between healthcare 7 Yes 102 70.3 professionals No 43 29.7 Monitoring 6 Subtotal 145 Review 1 Insufficient data 60 Management of complications 1 Total 205 Answers may be multiple n=36, not answered in 3

The Reviewers believed that there was room for improvement in 21% (39/190) of the patients treated in critical care (Table 7.57), the principal area for improvement being documentation (19/36) (Table 7.58).

Table 7.57 Room for improvement in critical care – Reviewers’ opinion Room for improvement Number of % patients Yes 39 20.5 No 151 79.5 Subtotal 190 Insufficient data 15 Total 205

93 Initial management of sepsis

Key Findings

• Half of the patients with sepsis were managed by acute • Following review antimicrobial therapy was modified in medical teams (270/539; 50%) 360/489 (73.6%) of patients in this group • The Reviewers considered that escalation/commencement • A pathogen was identified in 198/481 (41.2%) patients. of treatment was not timely in 98/504 (19.4%) patients The commonest pathogens were E coli, other coliforms • One in three patients (207/525; 39.4%) were started on and streptococcus a sepsis care bundle • Clinicians responsible for the patient considered that the choice of antimicrobial was not made in line with local • Management on a care bundle was associated with hospital policy in 193/593 (32.5%) cases fewer delays in the treatment of patients with sepsis • A source of infection was identified in 434/493 (88%) • 71/412 (17.2%) patients were not given or had delayed of cases and in 137/478 (28.7%) of these cases was the fluid administration source amenable to a procedure for control • There was room for improvement in the patient’s fluid • In 16/46 patients more could have been done to identify management in 203/447 (45.4%) cases the source • In 62.6% (226/361) of patients antimicrobials were • In patients in whom a source was amenable to control, administered within one hour of diagnosis that control was delayed in 55/129 (42.6%) • According to Reviewers, there was an avoidable delay in • Delay in source control affected the outcome in 33/47 the administration of antimicrobials in 114/391 (29.2%) patients patients • The Critical Care Outreach Team arrived promptly on • The Reviewers felt that the delay in the administration of 237/267 (88.8%) occasions antimicrobials affected the outcome in 44.3% (43/97) of those patients who did not receive antibiotics in a timely • Critical care services responded in an appropriate time manner frame in 222/238 (93.3%) referrals • The correct dose of antimicrobial was prescribed in 405/414 (97.8%) of patients • A microbiologist was consulted on the suitability of therapy in half (244/471; 51.8%) of the patients • Escalation was considered in 358/420 (85.2%) patients. De-escalation was considered in 289/389; 74.3%) and the duration of therapy considered in 329/413 (79.7%) patients

94 8

Complications of sepsis and discharge planning Back to contents

Survivors of sepsis are frequently left with a legacy of long- Table 8.1 Evidence of complications of sepsis at step term physical, neurological, psychological, and quality of down from Level 3 care (or if patient did not go to Level 3 care then at any time during the inpatient life impairments. Furthermore, beneficial interventions are episode) increasingly being identified to help patient management and reduce the prevalence and impact of these long-term Evidence Number of % patients complications. One in five patients in this cohort had evidence of complications following their episode of sepsis. Yes 141 32.6 The Reviewers considered that the majority of patients had No 292 67.4 complications identified in a timely fashion and adequately Subtotal 433 documented but there could have been improvement in the Insufficient data 118 manner in which they were appropriately treated. It was Total 551 also considered that rehabilitation was adequately planned. The commonest specialty referred to post-discharge was have experienced severe sepsis can suffer during their physiotherapy followed by occupational therapy and speech rehabilitation period. Any critical illness and time being and language therapy. treated in a critical care unit is already recognised as causing certain long-term problems. However, sepsis can cause On step down from critical care (or for patients who were additional physical and psychological problems which may not admitted to critical care, during the admission) there not become apparent for several weeks and these patients was evidence of complications of sepsis found in one third are more prone to recurring infections during the recovery of the study sample (141/433; 33%) (Table 8.1). period. The details of the complications in this group of patients and the timeliness in identifying them, the Post Sepsis Syndrome (PSS) is the term used to describe documentation and the appropriateness of treatment are the group of long-term problems that some patients who shown in Table 8.2.

Table 8.2 Complications related to sepsis

Number of patientsTimely identificationNo timely identificationInsufficient dataAdequate documentationNo adequate documentationInsufficient dataAppropriate treatmentNo appropriateInsufficient treatment data Worsened physical function 62 51 4 7 48 3 11 42 2 18 Worsening cognitive state 31 27 2 2 25 1 5 22 1 8 Post sepsis syndrome 13 4 4 5 4 2 7 3 1 9 Chronic pain 9 5 1 3 5 0 4 3 0 6 Post traumatic stress disorder 7 1 4 2 1 1 5 0 0 7 Amputation 6 2 0 4 2 0 4 2 0 4 Other 69 28 3 38 26 0 43 26 0 43 Answers may be multiple n=141

95 Complications of sepsis and discharge planning

Worsened physical function and cognitive state were the Table 8.4 Rehabilitation planning adequate for most common complications. Complications listed under patients discharged the ‘other’ category include acute kidney injury, cardiac Adequate rehabilitation plan Number of % arrest and hospital-acquired pneumonia (Table 8.3). patients Yes 284 96.6 Table 8.3 Other complications of sepsis No 10 3.4 Other complications Number of Subtotal 294 patients Insufficient data 97 Acute kidney injury/ impaired renal 11 function Total 391 Cardiac arrest 10 Hospital-acquired pneumonia 6 Table 8.5 Specialty referrals post discharge Further surgery 4 Referrals made Number of % Delirium 3 patients Wound problems 3 Physiotherapy 138 53.5 Fatigue 3 Occupational therapy 93 36.0 Further drainage required 3 Psychology 9 3.5 Tracheostomy 3 Specialist rehabilitation 28 10.9 Atrial fibrillation 2 Speech & language therapy 59 22.9 Depression 2 Other therapy 20 7.8 Fluid overload 2 No referral 3 1.2 Muscle weakness 2 Other 37 14.3 Weight loss 2 Answers may be multiple n=258; insufficient data in 133 Another infection 2 Table 8.6 Follow-up appointment post discharge Ischaemic bowel 2 Gastrointestinal bleed 2 Follow-up appointment with Number of % patients There were 391 patients in this study who were alive at Admitting physician 95 28.5 discharge, although Reviewers found evidence that 11 General practitioner 56 16.8 of these patients died after discharge. It was felt that the Admitting surgeon 71 21.3 rehabilitation planning for the patients who were alive at No follow-up appointment 61 18.3 discharge was appropriate in 284/294 (97%) cases Intensivist who cared for the 2 <1 (Table 8.4). patient Other 80 24.0 Referrals were most commonly made to physiotherapy No referral 3 <1 (138/158; 54%) but only made to psychology in 9/258 Other 37 11.1 (4%) cases (Table 8.5). Answers may be multiple n= 333; not answered in 58

Follow-up appointments were not common with only 95/333 The follow-up appointments, where made, were considered (29%) admitting physicians seeing these patients following by the Reviewers to be appropriate in 257/273 (94%) discharge from hospital (Table 8.6). More systematic follow- patients (Table 8.7); where it was considered inappropriate, up might have identified complications sooner and resulted in the reason given was mostly a lack of follow-up with the better post-discharge management of these patients. treating specialist physician or surgeon.

96 8

Table 8.7 Appropriate follow-up appointments Table 8.8 Evidence in the case notes that the GP was were made – Reviewers’ opinion informed of the admission – Reviewers’ opinion Appropriate follow-up Number of % GP was informed of the Number of % patients admission patients Yes 257 94.1 Yes 222 75.5 No 16 5.9 No 72 24.5 Subtotal 273 Subtotal 294 Insufficient data 118 Insufficient data 97 Total 391 Total 391

In order for patients to be adequately managed following Table 8.9 GP was given a copy of the discharge discharge from hospital the patients’ general practitioner (GP) summary – Reviewers’ opinion must be informed of the hospital admission, diagnosis and GP given copy of the Number of % post discharge management plan. It appears that for 1 in 4 discharge summary patients patients GPs were not informed of their admission (Table 8.8). Yes 222 97.8 However there was better compliance with the patient’s GP No 5 2.2 receiving a discharge summary (222/227; 97%). There was Subtotal 227 however, insufficient data to answer in 164 cases (Table 8.9). Insufficient data 164 Total 391 In the clinician questionnaire, the treating clinician was asked if sepsis was mentioned on the discharge summary (Table 8.10), and it was in just over half (264/490; 53%) of Table 8.10 Sepsis was mentioned on the discharge summary – Clinician questionnaire cases. Increasing this proportion could be an opportunity to improve feed-back to primary care using standard Sepsis mentioned Number of % patients terminology and raise awareness of sepsis throughout the pathway and the risk of repeat episodes of sepsis. Yes 264 53.9 No 226 46.1 The Reviewers were asked if, in their opinion, sufficient Subtotal 490 information was given to the patient, their relatives and Not applicable - still an inpatient 10 carers on discharge from hospital (Table 8.11). It was not Not answered 210 possible to ascertain whether this had occurred in many Total 710 patients. In several cases there was no evidence that patients, their relatives or carers had received either verbal or written information about the disease and its consequences.

Table 8.11 Sufficient information given to the patient/relatives/care giver – Reviewers’ opinion Sufficient information provided Patient % Relatives % Care giver % Yes 109 82.0 86 76.1 27 62.8 No 24 18.0 27 23.9 16 37.2 Subtotal 133 113 43 Insufficient data /not answered 245 267 295 Not applicable 13 11 53 Total 391 391 391

97 Complications of sepsis and discharge planning

One in five patients had evidence of complications at Table 8.14 Patient’s discharge from hospital was discharge (Table 8.12). The commonest being worsened delayed – Reviewers’ opinion physical function, which was identified in over half of Delay in discharge Number of % patients (38/71) (Table 8.13). patients Yes 68 19.3 Table 8.12 Complications present at discharge No 284 80.7 Complications Number of % Subtotal 352 patients Insufficient data 39 Yes 71 21.5 Total 391 No 260 78.5 Subtotal 331 Table 8.15 Length of delay to discharge Insufficient data 60 Total 391 Days Number of patients 0-2 days 4 Table 8.13 List of complications present at discharge 3-6 days 24 Complications at discharge Number of % 7-14 days 11 patients 15-30 days 13 Worsened physical function 38 53.5 > 30 days 4 Worsened cognitive state 14 19.7 Subtotal 56 Kidney injury/ impaired kidney 10 14.1 function Insufficient data 12 Post-sepsis syndrome 4 5.6 Total 68 Wound problems 4 5.6 Chronic pain 9 12.7 Table 8.16: Patient was readmitted Wound problems 3 4.2 Readmitted Number of % Amputation 6 8.5 patients Tracheostomy 3 4.2 Yes 31 10.1 Post traumatic stress disorder 7 9.9 No 275 89.9 Atrial fibrillation 3 4.2 Subtotal 306 Recurrence of sepsis 2 2.8 Insufficient data 85 Weight loss 2 2.8 Total 391 Depression 2 2.8 Muscle weakness 2 2.8 The treating clinician provided information regarding the Answers may be multiple n=71 functional status of patients on admission and at discharge. These data are shown in Figure 8.1. It can be seen that in In 19% of those patients with complications at discharge general, the worse a patient’s condition on admission, the the Reviewers considered that the complication had delayed worse their outcome at discharge. However, of the group the patients discharge from hospital (Table 8.14). The length who had no disability on admission, just over 20% had of the delay is shown in Table 8.15. moderate disability or worse at discharge (Figure 8.1).

98 8

Percentage of patients

100

90

80

70

60

50

40

30

20

10

0 No disability Slight disability Slight-moderate Moderate Moderate-severe Severe disability n=224 n=118 disability n=109 disability n=86 disability n=65 n=23 Functional status on admission

Functional status on discharge: Patient died Severe disability Moderate/severe disability Moderate disability Slight disability No disability

Figure 8.1 Functional status of patients at admission and at discharge – Clinician questionnaire

Number of patients

20

18

16

14

12

10

8

6

4

2

0 Re-infection New infection Complications of episode Other of sepsis Figure 8.2 Reason for readmission

Reviewers found evidence that one in ten patients (31/306; re-infection (Figure 8.2). The Reviewers were of the opinion 10%) were readmitted to hospital (Table 8.16). Of these that 5/23 of the patients who were readmitted could have patients one fifth had a new infection and another fifth had had this readmission prevented (data not shown).

99 Complications of sepsis and discharge planning

End of life care Table 8.17 Treatment withdrawal in patients who died Medical treatment should only be withdrawn on clinical Treatment withdrawn Number of % grounds because the treatment will not benefit the patient patients or the expected benefits are outweighed by the burdens Yes 94 66.2 of treatment. There is a need to know, where possible, No 48 33.8 the wishes of the patient. If the patient is not competent Subtotal 142 to communicate their wishes, their family and friends Insufficient data 6 should be consulted. Attempts must be made to discover the patient’s wishes in this situation. Advance directives, Total 148 if available and relevant, may be helpful.60 Many units consider it good practice to involve two senior clinicians in The decision to withdraw treatment was made by a clinician the decision making process when considering withdrawal of suitable seniority in 87/90 cases, by more than one of treatment. clinician in 73/82 cases. The decision was discussed with the patients’ relatives in 86/89 cases, the patient in 14/42 Treatment was withdrawn from 66% (94/142) of patients cases and a patient advocate in 16/27 cases (Table 8.18). who died during the admission (Table 8.17). The Reviewers The reason for the “discussion with patient category” being considered this appropriate in all but three. so low is believed to be because a number of very unwell patients in Level 3 care would have been sedated when these decisions were made. Table 8.18 Decisions about withdrawal of treatment Treatment withdrawal discussions: Yes No Subtotal Insufficient Total data/ Not applicable Made by clinician of appropriate seniority 87 3 90 4 94 Made by more than one clinician 73 9 82 12 94 Discussed with the patient 14 28 42 52 94 Discussed with relatives 86 3 89 5 94 Discussed with patient advocate 16 11 27 67 94

Table 8.19 Patient was on an end of life care Thirty-two patients were placed on end of life care pathway pathways (Table 8.19). The Reviewers considered this to be Patient was on an end of life Number of % appropriate in all cases (Table 8.20). care pathway patients Yes 32 23.7 No 103 76.3 Subtotal 135 Insufficient data 13 Total 148

100 8

Table 8.20 Decisions about end of life care for patients on an end of life care pathway End of life care Yes No Subtotal Insufficient Total data/ Not applicable Made by clinician of appropriate seniority 30 0 30 2 32 Made by more than one clinician 23 4 27 5 32 Discussed with the patient 7 8 15 17 32 Discussed with relatives 26 2 28 4 32 Discussed with patient advocate 4 9 13 19 32

Table 8.21 Involvement of a palliative care team Table 8.22 Palliative care team should have been Palliative care team involved Number of % involved – Reviewers’ opinion patients Palliative care team should have been Number of Yes 20 27.8 involved patients No 52 72.2 Yes 6 Subtotal 72 No 29 Not applicable 469 Subtotal 35 Insufficient data 10 Insufficient data 17 Total 551 Total 52

In only 20 patients were palliative care teams involved (Table Table 8.23 Outcome of patients in the study at 30 8.21). The Reviewers considered that palliative care should days post diagnosis of sepsis have been involved in six further patients (Table 8.22). Outcome at 30 days Number of % patients There were 148 patients in the study case review cohort Discharged alive 380 69.1 who died during their admission. A further 11 patients Inpatient at 30 days 11 2.0 were identified who died following discharge from hospital, Died during admission 148 26.9 making a total of 159 patients who did not survive (159/551; Died after discharge 11 2.0 29%). There were 11 patients who were still an inpatient at Subtotal 550 the time of data collection and one patient in the case review Unknown mortality status 1 cohort for whom there was insufficient information available to establish if they survived (Table 8.23). Total 551

101 Complications of sepsis and discharge planning

Number of patients Male Female

35

30

25

20

15

10

5

0 17-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 >90

Age range (years) Figure 8.3 Age and gender of the patients in this study who died

Figure 8.3 shows the age profile of the patients in the The Reviewers were of the opinion that of those cases for cohort who died. The age distribution of this group is which sepsis was not mentioned on the death certificate as similar to the whole cohort but with a greater proportion of a cause of death, that it should have been in 48/59 cases patients in the ninth decade of life. (Table 8.26). Information from death certificates is used to measure the relative contributions of different diseases An autopsy was only performed in 12% (15/124) of to mortality. Statistical information from death certificates those patients who died (Table 8.24) and sepsis was only is important for monitoring the health of the population, mentioned in the causes of death on 40% (42/103) of death evaluating public health interventions, recognising priorities certificates (Table 8.25). for medical research and health services, planning health services, and assessing the effectiveness of healthcare. The Table 8.24 Autopsy performed treating clinician completing the questionnaire felt that the Autopsy performed Number of % death was preventable in 10/192 (5%) of cases (Table 8.27). patients In 6/10 of these cases, the reasons given related to delay in Yes 15 12.1 diagnosis of sepsis and/or antimicrobial therapy. No 109 87.9 Subtotal 124 Table 8.26 Sepsis should have been recorded on the death certificate – Reviewers’ opinion Insufficient data 36 Sepsis should have been Number of % Total 159 recorded patients Yes 48 81.4 Table 8.25 Sepsis recorded on the death certificate No 11 18.6 Sepsis recorded Number of % Subtotal 59 patients Insufficient data 2 Yes 42 40.8 Total 61 No 61 59.2 Subtotal 103 Insufficient data 46 Total 149

102 8

Nearly two-thirds of patients who died with sepsis (where it Of the 48 cases where sepsis was not mentioned on the could be identified by the treating clinician) were discussed death certificate but should have been, according to the at a morbidity and mortality (M&M) meeting (Table 8.28). Reviewers, 18 cases were discussed at an M&M meeting according to the treating clinician. M&M meetings have Table 8.27 Preventable death in the view of the potential to provide accountability and the necessary the clinician caring for the patient – Clinician improvement measures required for patient safety as well questionnaire as professional learning. Studies have shown that for M&M Death preventable Number of % meetings to facilitate improvement and be more than a patients forum for peer review, they need to be structured and Yes 10 5.2 systematic in reviewing and discussing deaths, directing No 151 78.6 discussions towards improving system and process 65 Unknown 31 16.1 variations. Total 192

Table 8.28 Case discussed at a morbidity and mortality meeting – Clinician questionnaire Case discussed Number of % patients Yes 69 63.9 No 39 36.1 Subtotal 108 Unknown 74 Not answered 10 Total 192

103 Complications of sepsis and discharge planning

Key Findings

• 71/331 (21.5%) patients had evidence of complications • The discharge was delayed in 68/352 (19.3%) cases at discharge • The decision to withdraw treatment was made by a • The most common complication was worsened physical clinician of suitable seniority in 87/90 cases function (38/71; 53.5%) • For those placed on end of life care pathways (32/135; • 31/306 (10.1%) patients were readmitted to hospital 23.7%), 100% were found to be appropriate following an episode of sepsis • An autopsy was performed in 15/124 (12.1%) patients • No follow-up appointment was made for 61/333 who died (18.3%) patients • Sepsis was included on the death certificate in 42/103 • According to Reviewers there was no evidence that the (40.8%) patients who died. Of those where it was not GP was informed of the admission in 72/294 (24.5%) included Reviewers considered that it should have been cases in 48/59 (2 not answered) • Sepsis was not mentioned on the discharge summary in • Cases were documented as being discussed at M&M 226/490 (46.1%) of cases meetings in 69/108 (63.9%) patients who died • There was evidence of insufficient information being given to patients on discharge in 24/133 cases

104 9

Overall quality of care Back to contents

The Reviewers were asked to comment on the overall quality Table 9.1 Overall quality of care as rated by the of care received by patients in the study. Just over one third Reviewers of the study population were considered to have received Overall quality of care Number of % good care during their admission. Most commonly in the patients group of patients who were judged to have received less Good practice 198 36.5 than good care, it was considered that there was room for Room for improvement (clinical) 149 27.4 improvement in clinical aspects of their care rather than Room for improvement 39 7.2 organisational factors. This suggests that the deficiencies are (organisational) more in the management, awareness and decision making Room for improvement (both) 123 22.7 of the doctors and nurses caring for these patients rather Less than satisfactory 34 6.3 than systematic deficiencies in process or the organisation of Subtotal 543 services or equipment (Table 9.1, Figure 9.1). Insufficient data 8 Total 551

Percentage

45

40 198 35

30 149

25 123

20

15

10 39 34 5

0 Good practice Room for Room for Room for Less than improvement improvement improvement satisfactory (clinical) (organisational) (clinical and organisational)

Figure 9.1 Overall quality of care

105 106 Recommendations Back to contents

1. All hospitals should have a formal protocol for the early suspected. This will aid the recognition of the severity identification and immediate management of patients of sepsis and can be used to prioritise urgency of care. with sepsis. The protocol should be easily available (General Practitioners, Ambulance Trusts, Health Boards, to all clinical staff, who should receive training in its NHSE, Clinical Directors, Royal Colleges) use. Compliance with the protocol should be regularly audited. This protocol should be updated in line with 6. Primary care providers should ensure that robust safety changes to national and international guidelines and netting arrangements are in place for those patients local antimicrobial policies. (Medical Directors) who are suspected to be at risk of sepsis. (General Practitioners) 2. Training in the recognition and management of sepsis in primary and secondary care should be included 7. To facilitate the transition from primary to secondary in educational materials for healthcare professionals care, a standard method of referral should be undertaking new posts. Where appropriate this introduced in primary care for patients who are in need training should include the use of a standardised of a hospital admission for, or thought to be at risk of, hospital protocol (Medical Directors, Nursing Directors, sepsis. This should include a full set of observations/ Postgraduate Deaneries, Health Education England, vital signs/risks/relevant history (such as previous sepsis) Royal Colleges) and any early warning scores used. (Primary Care Practitioners, Commissioners) 3. A Clinical Lead in sepsis should be appointed in every Trust/Health Board to champion best practice and take 8. On arrival in the emergency department a full set of responsibility for the clinical governance of patients vital signs, as stated in the Royal College of Emergency with sepsis. This Lead should also work closely with Medicine standards for sepsis and septic shock should those responsible for antimicrobial stewardship in their be undertaken. (Emergency Medicine Physicians, Clinical hospital(s). (Medical Directors, Nursing Directors, Trust Directors, Nursing Directors) Chief Executives) 9. Where sepsis is suspected, early consideration should be 4. Trusts/Health Boards should use a standardised sepsis given to the likely source of infection and the ongoing proforma to aid the identification, coding, treatment management plan recorded. Once identified, control of and ongoing management of patients with sepsis the source of infection should be undertaken as soon (some examples are available at sepsistrust.org and as possible. Appropriate staffing and hospital facilities survivingsepsis.org). To ensure continuity of care, this (including theatre/interventional radiology) should be proforma should be compatible, where possible with available to allow this to occur. (Medical Directors, any similar proforma or system used in primary care Clinical Directors) and should permit the data to be shared electronically. (Medical Directors, Primary Care Practitioners, 10. The importance of early identification and control of the Commissioners) source of sepsis should be emphasised to all clinicians, and be reinforced in any future guidelines or tools for 5. An early warning score, such as the National Early the management of sepsis. (International Sepsis Forum, Warning Score (NEWS) should be used in both primary UK Sepsis Trust, NICE, Health Education England, care and secondary care for patients where sepsis is Postgraduate Deaneries, Royal Colleges)

107 Recommendations

11. In line with previous NCEPOD and other national 16. A booklet that provides patients and their relatives with reports’ recommendations on recognising and caring easy to understand information on the recognition of for the acutely deteriorating patients, hospitals should sepsis, its long-term complications, recovery and risk ensure that their staffing and resources enable: of recurrence should be available from all healthcare a. All acutely ill patients to be reviewed by a providers and be provided to patients with sepsis at consultant within the recommended national discharge from hospital. Some examples can be found timeframes (max of 14 hours after admission) at the UK Sepsis Trust (sepsistrust.org) and ICU Steps b. Formal arrangements for handover (icusteps.org). (Medical Directors, Commissioners) c. Access to critical care facilities if escalation is required; and 17. As for all acutely ill patients who are admitted to critical d. Hospitals with critical care facilities to provide care, a follow-up service for patients with sepsis should a Critical Care Outreach service (or equivalent) be provided by the hospital which includes support and 24/7. (Medical Directors, Nursing Directors, rehabilitation services, as recommended in NICE Clinical Commissioners) Guideline 83 and the Faculty of Intensive Care Medicine and Intensive Care Society Guidelines for the Provision 12. All patients diagnosed with sepsis should benefit from of Intensive Care Services (GPICS). (Medical Directors, management on a care bundle as part of their care Clinical Directors, Sepsis Leads) pathway. The implementation of this bundle should be audited and reported on regularly. Trusts/Health Boards 18. All patients discharged following a diagnosis of sepsis should aim to reach 100% compliance and this should should have sepsis recorded on the discharge summary be encouraged by local and national commissioning provided to the general practitioner so that it can arrangements. (Medical Directors, Clinical Directors, be recorded in the patient’s GP record. (All Hospitals Commissioners) Doctors, General Practitioners)

13. For any invasive procedure a surgical site bundle should 19. For patients who die with sepsis, the care be employed as specified in NICE Clinical Guideline 74. provided should always be discussed at a hospital (Medical Directors, Clinical Directors) multidisciplinary mortality meeting to encourage learning, and, where the source of sepsis has not been 14. All healthcare providers should ensure that identified, an autopsy should be undertaken. (Medical antimicrobial policies are in place including prescription, Directors, Clinical Directors, Clinical Governance Leads, review and administration of antimicrobials as part of Sepsis Leads, All Clinical Staff) an antimicrobial stewardship process. These policies must be accessible, adhered to and frequently reviewed 20. When diagnosed, sepsis should always be included with training provided in their use. (Medical Directors, on the death certificate, in addition to the underlying Commissioners, General Practitioners, Postgraduate source of infection. (All Doctors including Sepsis Leads) Deaneries, Health Education England) 21. The use of national coding for sepsis must be improved 15. There should be senior microbiology input into the in order to aid clinical audit, national reporting and management of all patients identified with sepsis. This shared learning. Use of a standardised proforma as input should be available 24/7 and sought early in the described in recommendation 4 should help improve care pathway. (Medical Directors, Sepsis Leads, Clinical this process, and may help in the development of a Directors) national registry. (Chief Executives, Medical Directors, Clinical Governance Leads, Sepsis Leads)

108 Summary Back to contents

This study set out to identify and explore avoidable and procedures. In 10/88 patients with hospital-acquired remediable factors in the process of care for patients with infection, the Reviewers felt that the infection was known or suspected sepsis. From the cases identified, the preventable. Reviewers were able to assess 551 cases. Of these, 54 sets of general practitioner (GP) notes were received and suitable The Reviewers considered that there was a delay in for review. identifying sepsis in 182/505 (36%) cases, severe sepsis in 167/324 (51%) and septic shock in 63/193 (32%), and This study confirmed that there is huge variability in identified that good documentation of sepsis was associated the clinical presentation of sepsis. Patients seen in the with more timely diagnosis. Despite the presence of community present diagnostic dilemmas and whilst the protocols, investigations considered essential in the diagnosis difficulty is recognised, it was of note that there was poor of sepsis were missed in 39% of patients and delayed in recording of clinical observations by primary and secondary 39%. Management on a care bundle reduced delays in the care providers that may have assisted with both the treatment of patients with sepsis. However, only 39.4% of immediate management and handover between primary patients were started on a sepsis care bundle. This study and secondary care. Half of the patients referred to hospital highlights the absolute requirement for hospitals accepting by GPs had no referral letter. The use of pre-alerts was only emergency admission to have a formal protocol for the early apparent in 8 patients, although 50% of hospitals reported identification and immediate management of patients with they were available, and in the Emergency Department (ED) sepsis. Only 55/215 (25.6%) acute hospitals used standard 40% of patients did not have a timely review by a senior proformas to identify and monitor patients with sepsis, and clinician. less than half (90/204; 44%) audited the timely treatment of severe sepsis against their own protocols. It is recognised The importance of source control is often overlooked and that if clinical management is to improve, clinical leadership it was noted that a possible source of infection was only is important. However, only half of the hospitals in the study recorded at triage in 46% of patients admitted via the ED. (166/322; 52%) had appointed a lead clinician for sepsis. And in those patients in whom a source was amenable to control, that control was delayed in 43% of cases which This is a group of patients who benefit from the use of could have affected the outcome in 26/41 patients in the antimicrobials, but with the current awareness of over use view of the case Reviewers. of antimicrobials, antimicrobial stewardship is important; not only in the management of sepsis but also the in Following admission to hospital, 20% of the patients in broader environment of healthcare. It was of note that a this study were not seen by a consultant within 14 hours. microbiologist was consulted on the suitability of therapy In view of the fact that 61.5% patients had changes made in only 52% of patients. This was also reflected in the to their care following consultant review, it is paramount organisational data. Senior microbiological input is essential that the resources are in place to ensure prompt consultant in the management of patients with sepsis to aid the review. appropriateness of antimicrobial usage.

One quarter of the patients in this study acquired their Morbidity following sepsis is common and 22% patients infection whilst in hospital. In half of these patients the had evidence of complications at discharge. There was little infection was diagnosed following an invasive procedure. evidence of information being given to sepsis patients on the A surgical site bundle was only utilised in 43/73 invasive disease and its consequences.

109 summary

For those patients who died, an autopsy was only performed in 12.1% of cases, sepsis was only included on the death certificate in 40.8% and only 63.8% of cases were discussed at mortality and morbidity reviews, missing opportunities to learn from the care provided.

Throughout the patient pathway areas for improvement were identified and the Reviewers were of the opinion that good care was delivered in only 36% of cases. Early recognition, better documentation and prompt treatment of sepsis would all lead to improved care for this group of patients. Using the word ‘sepsis’ as soon as it is considered would also raise awareness amongst healthcare professionals and patients.

110 References Back to contents

1. Dellinger RP, Levy MM and Rhodes A et al. Surviving 10. SEPSIS Management in Scotland: A Report by the Sepsis Campaign: international guidelines for Scottish Trauma Audit Group. 2010. management of severe sepsis and septic shock: 2012. http://www.stag.scot.nhs.uk/SEPSIS/Main.html Crit Care Med. 2013; 41: 580-637 11. Levy MM, Dellinger RP, Townsend SR et al. The Surviving 2. International Sepsis Forum. http:// Sepsis Campaign: results of an international guideline internationalsepsisforum.com based performance improvement program targeting severe sepsis. Crit Care Med. 2010; 38: 367-74 3. Esteban A, Frutos-Vivar F and Ferguson ND et al. Sepsis incidence and outcome: Contrasting the intensive care 12. National Institute for Health and Care Excellence unit with the hospital ward. Crit Care Med. 2007; 35(5): (NG15). Antimicrobial stewardship: systems and 1284-1289. processes for effective antimicrobial medicine use. 2015 www.nice.org.uk/guidance/ng15 4. The College of Emergency Medicine. Sepsis Audit. 2011 http://www.rcem.ac.uk/Shop-Floor/Clinical%20Audit/ 13. NHSE anti microbial patient safety alert – addressing Previous%20Audits antimicrobial resistance through implementation of an antimicrobial stewardship programme. August 2015. 5. The College of Emergency Medicine. Sepsis and Septic NHS England http://www.england.nhs.uk/wp-content/ Shock Audit. 2014 http://www.rcem.ac.uk/Shop-Floor/ uploads/2015/08/psa-amr-stewardship-prog.pdf Clinical%20Audit/Previous%20Audits 14. Shahin J, Harrison DA and Rowan KM. Relation 6. National Confidential Enquiry into Patient Outcome and between volume and outcome for patients with severe Death (NCEPOD). An acute problem? (2005) sepsis in United Kingdom: retrospective cohort study. http://www.ncepod.org.uk/2005aap.htm BMJ. 2012; 344: e3394–4.

7. National Institute for Health and Care Excellence 15. The forward view into action: planning for 2015/16. (CG50). Acutely ill patients in hospital: Recognition of NHS England and response to acute illness in adults in hospital. 2007 http://www.england.nhs.uk/wp-content/ https://www.nice.org.uk/guidance/cg50 uploads/2014/12/forward-view-plning.pdf

8. The fifth report from the Patient Safety Observatory. 16. Commissioning for Quality and Innovation (CQUIN) Safer care for the acutely ill patient: learning from Guidance for 2015/16 http://www.england.nhs.uk/wp- serious incidents. NPSA 2007 content/uploads/2015/03/9-cquin-guid-2015-16.pdf

9. Royal College of Physicians. National Early Warning 17. NICE clinical guideline development group for sepsis Score (NEWS): Standardising the assessment of acute https://www.nice.org.uk/guidance/indevelopment/gid- illness severity in the NHS. Report of a working party. cgwave0686 London: RCP, 2012. https://www.rcplondon.ac.uk/sites/ default/files/documents/national-early-warning-score- standardising-assessment-acute-illness-severity-nhs.pdf

111 References

18. Welsh Intensive Care Society. The size of sepsis in Wales: 26. Bochud PY, Bonten M and Marchetti O et al. point prevalence study of sepsis in the acute hospital Antimicrobial therapy for patients with severe sepsis v3.1. 2015 and septic shock: an evidence-based review. Crit Care http://welshintensivecaresociety.org/wp-content/ Med. 2004; 32: S495-S512 uploads/2014/03/Size-of-sepsis-in-Wales-protocol- version-3.1.pdf 27. Robb E, Jarman B and Suntharalingam G et al. Using care bundles to reduce in-hospital mortality: 19. Szakmany T, Ellis G and Lundin R et al. Size of sepsis in quantitative survey BMJ. 2010;340: c1234 Wales: feasibility pilot. Crit Care Med. 2014; 42(12): pA1446 28. Rivers E, Nguyen B and Havstad S et al. Early Goal- Directed Therapy Collaborative Group Early Goal- 20. Knight M, Kenyon S, Brocklehurst P, Neilson J, Directed Therapy in the Treatment of Severe Sepsis and Shakespeare J, Kurinczuk JJ (Eds.) on behalf of Septic Shock. N Engl J Med. 2001; 345: 1368-1377. MBRRACEUK. Saving Lives, Improving Mothers’ Care - Lessons learned to inform future maternity care from 29. National Confidential Enquiry into Patient Outcome and the UK and Ireland Confidential Enquiries into Maternal Death (NCEPOD). Cardiac Arrest Procedures: Time to Deaths and Morbidity 2009–12. Oxford: National Intervene (2012). Perinatal Epidemiology Unit, University of Oxford 2014 http://www.ncepod.org.uk/2012cap.htm https://www.npeu.ox.ac.uk/downloads/files/mbrrace-uk/ reports/Saving%20Lives%20Improving%20Mothers%20 30. Iwashyna TJ, Dmith DM and Langa KM. Long-term Care%20report%202014%20Full.pdf cognitive impairment and functional disability among survivors of severe sepsis. Jama. 2010; 304 (16): 1787- 21. Parliamentary and Health Service Ombudsman. TIME TO 1794 ACT Severe sepsis: rapid diagnosis and treatment saves lives. 2013 31. National Institute for Health and Care Excellence http://www.ombudsman.org.uk/__data/assets/pdf_ (NG15). Rehabilitation after critical illness. 2009. https:// file/0004/22666/FINAL_Sepsis_Report_web.pdf www.nice.org.uk/guidance/cg83

22. All Party Parliamentary Group (APPG) on sepsis. October 32. Kahn JM, Benson NM, Appleby D, Carson SS, Iwashyna 2015 TJ. Long-term acute care hospital utilization after critical illness. JAMA: the journal of the American Medical 23. Daniels R, Nutbeam T and McNamara G et al. The Association. 2010; 303(22): 2253-2259 sepsis six and the severe sepsis resuscitation bundle: a prospective observational cohort study. Emerg Med J. 33. Statistics on Obesity, Physical Activity and Diet: England 2011; 28: 507-512 2014. HSCIC

24. Paul M, Shani V, Muchtar E et al. Systematic review 34. Trenkwalder P, Ruland D and Stender M et al. and meta-analysis of the efficacy of appropriate Prevalence, awareness, treatment and control of empiric antibiotic therapy for sepsis. Antimicrob Agents hypertension in a population over the age of 65 years: Chemother. 2010; 54: 4851-63 results from the Starnberg Study on Epidemiology of Patkinsonism and Hyperension in the Elderly (STEPHY). J 25. Kumar A, Roberts D, Wood KE et al. Duration of Hypertens. 1994; 12(6): 709-716 hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human 35. Ash.org.uk septic shock. Crit Care Med. 2006; 34: 1589-96

112 36. Almond S, Mant D and Thompson M. Diagnostic safety- 45. Rotter T, Kinsman L and James EL et al. Cochrane netting. Br J Gen Pract. 2009; 59(568): 872-874 Effective Practice and Organisation of Care Group. Clinical pathways: effects on professional practice, 37. RCP Acute care toolkit4: Delivering a 12-hour, 7-day patient outcomes, length of stay and hospital costs. consultant presence on the acute medical unit. 2012. Cochrane Database Syst Rev. 2010; 3(3): CD006632 https://www.rcplondon.ac.uk/resources/acute-care- toolkit-4-delivering-12-hour-7-day-consultant-presence- 46. Madhusudan P, Vijayaraghavan BKT and Cove ME. acute-medical-unit Fluid resuscitation in sepsis: reexamining the Paradigm. BioMed Research International. 2014; 2014, Article ID 38. NHS England. 2013. NHS Services, Seven Days a Week 984082 Forum. Everyone Counts: Planning for Patients 2013/14 to 2018/19. http://www.england.nhs.uk/wp-content/ 47. Martin GS. Sepsis, severe sepsis and septic shock: uploads/2013/12/5yr-strat-plann-guid-wa.pdf changes in incidence, pathogens and outcomes. Expert Rev Anti Infect Ther. 2012; 10(6): 701–706 39. Academy of Medical Royal Colleges. The Benefits of Consultant Delivered Care. 2012. 48. Kumar A, Roberts D and Wood KE et al. Duration of http://www.aomrc.org.uk/doc_view/9450-the-benefits- hypotension before initiation of effective antimicrobial of-consultant-delivered-care therapy is the critical determinant of survival in human septic shock. Crit Care Med. 2006; 34(6): 1589-1596 40. Public Health England. English National Point Prevalence Survey on Healthcare-associated Infections and 49. National Emergency Laparotomy Audit. www.nela.org.uk Antimicrobial Use. 2011 https://www.gov.uk/government/uploads/system/ 50. Iregui M, Ward S and Sherman G et al. Clinical uploads/attachment_data/file/331871/English_ importance of delays in the initiation of appropriate National_Point_Prevalence_Survey_on_Healthcare_ antibiotic treatment for ventilator-associated associated_Infections_and_Antimicrobial_Use_2011.pdf pneumonia. Chest. 2002; 122(1): 262-268

41. Levy M M et al. Surviving Sepsis Campaign: Association 51. Zilberberg MD, Shorr AF and Micek MT et al. Between Performance Metrics and Outcomes in a 7.5- Antimicrobial therapy escalation and hospital Year Study. Crit Care Medicine. 2015; 43(1): p3–12. mortality among patients with health-care associated pneumonia: a single-center experience. Chest. 2008; 42. National Institute for Health and Care Excellence 134(5): 963-968 (CG74). Surgical site infection: Prevention and treatment of surgical site infection. 2008 52. Shorr AF, Micek ST and Welch EC et al. Inappropriate http://www.nice.org.uk/guidance/cg74 antibiotic therapy in Gram-negative sepsis increases hospital length of stay. Crit Care Med. 2011; 39(1): 43. Kaukonen KM, Bailey M and Pilcher D et al. Systemic 46-51 inflammatory response syndrome criteria in defining severe sepsis. N Engl J Med. 2015;372(17):1629-38 53. Kollef MH, Sherman G and Ward S et al. Inadequate antimicrobial treatment of infections: a risk factor for 44. Laudermilch DJ et al. Lack of emergency medical hospital mortality among critically ill patients. Chest. services documentation is associated with poor patient 1999; 115(2): 462-474 outcomes: A validation of audit filters for pre-hospital trauma care. J Am Coll Surg. 2010; 210(2): 220-227

113 References

54. Garnacho-Montero J, Garcia-Garmendia JL and Barrero- 58. Hillman K. Critical care without walls. Curr Opin Crit Almodovar A et al. Impact of adequate empirical Care. 2002; 8(6): 594-9 antibiotic therapy on the outcome of patients admitted to the intensive care unit with sepsis. Crit Care Med. 59. Priestley G, Watson W and Rashidian A et al. 2003; 31(12): 2742-2751 Introducing Critical Care Outreach: a ward-randomised trial of phased introduction in a general hospital. 55. Harbarth S, Garbino J and Pugin J et al. Inappropriate Intensive Care Medicine. 2004; 30(7): 1398–404 initial antimicrobial therapy and its effect on survival in a clinical trial of immunomodulating therapy for severe 60. The Intensive Care Society. Guidelines for limitations of sepsis. Am J Med. 2003; 115(7): 529-535 treatment for adults requiring intensive care. 2003 http://www.ics.ac.uk/EasysiteWeb/getresource. 56. Ashira-Oredope D, Sharland M and Charani E et al. axd?AssetID=469&type Improving the quality of antibiotic prescribing in the NHS by developing a new Antimicrobial stewardship 61. Batalden PB, Davidoff F. What is “quality improvement” programme: Start Smart-Then Focus. J. Antimicrob. and how can it transform healthcare? Qual Saf Health Chemother. 2012; 67 (suppl 1): i51-i63 Care. 2007; 16(1): 2-3

57. Jimenez MF, Marshall JC. Source control in the management of sepsis. Intensive Care Med. 2001; 27 (suppl 1): s49-62

114 Appendices Back to contents

Appendix 1 – Glossary

Acute Medical Unit (AMU) This is the first point of entry for patients referred to hospital as emergencies by their GP and those requiring admission from the Emergency Department.

Antimicrobial An antimicrobial kills microorganisms or inhibits their growth. Antimicrobial medicines can be grouped according to the microorganisms they act primarily against. For example, antibacterials are used against bacteria and antifungals are used against fungi.

Antimicrobial stewardship This refers to co-ordinated interventions designed to improve and measure the appropriate use of antimicrobials by promoting the selection of the optimal antimicrobial drug regimen, dose, duration of therapy, and route of administration.

AVPU The AVPU scale (an acronym from "alert, voice, pain, unresponsive") is a system by which a first aider, ambulance crew or healthcare professional can measure and record a patient's responsiveness, indicating their level of consciousness.

Candida Candida is a type of yeast and is the most common cause of fungal infections.

Care bundle A set of interventions that, when used together, improve patient outcomes.

Critical Care Outreach / Multidisciplinary teams, consisting of staff trained in intensive care who review acutely ill Rapid Response /Medical patients. Emergency Teams/Service

CDU/MAU The Clinical Decisions Unit (CDU) deals with all acute medical problems.

Coliform bacteria A broad class of bacteria found in our environment, including the faeces of humans and animals.

Critical care /ICU - Level 3/ Intensive Care (Level 3) or High Dependency Care (Level 2). Specialist hospital wards HDU - Level2 providing detailed care for very sick patients.

Cultures The growth of bacteria in a laboratory environment.

Early warning score/EWS/ This is a guide used by medical services to quickly determine the degree of illness of NEWS a patient. It is based on data derived from four physiological readings (systolic blood pressure, heart rate, respiratory rate, body temperature) and one observation (level of consciousness, AVPU). The ‘N’ stands for National.

Functional status ranking Slight disability: generally able to carry out activities unaided but may require assistance with certain tasks; moderate disability: Requiring some help but able to walk without assistance; moderate to severe disability: Unable to walk without assistance and unable to attend to own bodily needs without assistance; severe disability: Bedridden, incontinent and requiring constant nursing care and attention.

Glasgow Coma Scale This is a neurological scale to record the conscious state of a person – from 3 (indicating (GCS) deep unconsciousness) to 15.

115 Appendices

Gram-negative Bacteria that do not retain the crystal violet stain used in the Gram staining method of bacterial differentiation, making positive identification possible. E.g. Staphylocuccus aureus

Gram-positive Bacteria that take up the crystal violet stain used in the test, and then appear to be purple-coloured when seen through a microscope. E.g. Streptoccus pneumoniae

ICNARC Intensive Care National Audit & Research Centre

Inotropes An agent that alters the force or energy of muscular contractions.

MBRRACE-UK Mothers and Babies: Reducing Risk through Audits and Confidential Enquiries in the UK.

Post Sepsis Syndrome Physical and/or long term effect such as insomnia, fatigue or joint pain (PSS)

Procalcatoinin This is a marker of inflammatory response.

Sepsis A common and potentially life threatening condition triggered by infection. Sepsis is defined as infection plus systemic manifestations of infection. For the purposes of this study, the definition from the surviving sepsis campaign has been used: the presence of infection plus systemic manifestations of infection (see page 13)

Sepsis six The Sepsis Six is the name given to a bundle of medical therapies designed to reduce the mortality of patients with sepsis - three diagnostic and three therapeutic steps – all to be delivered within one hour of the initial diagnosis of sepsis.

1 Deliver high-flow oxygen. 2 Take blood cultures. 3 Administer empiric intravenous antibiotics. 4 Measure serum lactate and send full blood count. 5 Start intravenous fluid resuscitation. 6 Commence accurate urine output measurement.

Septic shock A life-threatening condition that happens when blood pressure drops to a dangerously low level after an infection.

Serious incident Adverse events, where the consequences to patients, families and carers, staff or organisations are so significant or the potential for learning is so great, that a heightened level of response is justified.

Severe sepsis Sepsis which is complicated by acute organ dysfunction.

Systemic inflammatory An inflammatory state affecting the whole body, frequently a response of the immune response syndrome (SIRS) system to infection.

Staphylococcus A type of Gram-positive bacteria.

Streptococcus A group of bacteria.

Track and trigger A system which uses periodic observations of basic vital signs (heart rate, blood pressure, etc.) together with pre-determined criteria to ensure timely recognition of deteriorating patients and to trigger a request for more experienced staff, usually the critical care outreach service.

Vasopressors A powerful class of drugs that induce vasoconstriction and thereby elevate mean arterial pressure (MAP). Vasopressors differ from inotropes, which increase cardiac contractility; however, many drugs have both vasopressor and inotropic effects. E.g. Noradrenaline

116 Appendix 2 – Existing sepsis information, templates and tools

http://sepsistrust.org/professional/professional-resources/ http://sepsistrust.org/clinical-toolkit/ http://sepsistrust.org/professional/educational-tools/

http://www.rcem.ac.uk/Shop-Floor/Clinical%20Standards/Sepsis Standards 1) Temperature, pulse rate, respiratory rate, blood pressure, mental status (AVPU or GCS) and capillary blood glucose on arrival. 2) Senior EM a ssessment of patient within 60mins of arrival. 3) High flow O2 via non-re- breathe mask was initiated (unless there is a documented reason to the contrary) before leaving the ED. 4) Serum lactate measured before leaving the ED. 5) Blood cultures obtained before leaving the ED. 6) Fluids - first intravenous crystalloid fluid bolus (up to 20mls/kg given: 75% within 1 hour of arrival; 100% before leaving the ED. 7) Antibiotics administered: 50% within 1 hour of arrival; 100% before leaving the ED. 8) Urine output measurements instituted before leaving the ED.

http://www.stag.scot.nhs.uk/SEPSIS/Forms.html

https://www.nice.org.uk/guidance/gid-cgwave0686/documents/sepsis-the- recognition-diagnosis-and-management-of-severe-sepsis-scope-consultation

https://www.rcplondon.ac.uk/resources/national-early-warning-score-news

117 Appendices

Appendix 3 – The role and structure of NCEPOD

The National Confidential Enquiry into Patient Outcome and Death (NCEPOD) is an independent body to which a corporate commitment has been made by the Medical and Surgical Colleges, Associations and Faculties related to its area of activity. Each of these bodies nominates members on to NCEPOD’s Steering Group.

Steering Group as at 24th November 2015 Dr A Hartle Association of Anaesthetists of Great Britain and Ireland Mr F Smith Association of Surgeons of Great Britain and Ireland Mr K Altman Faculty of Dental Surgery, Royal College of Surgeons of England Vacancy Faculty of Public Health Medicine Mr S Barasi Lay Representative Ms S Payne Lay Representative Dr J Fazackerley Royal College of Anaesthetists Dr A Batchelor Royal College of Anaesthetists Dr C Mann Royal College of Emergency Medicine Dr D Cox Royal College of General Practitioners Mrs J Greaves Royal College of Nursing Dr E Morris Royal College of Obstetricians and Gynaecologists Mr W Karwatowski Royal College of Ophthalmologists Dr I Doughty Royal College of Paediatrics and Child Health Dr M Osborn Royal College of Pathologists Mr M McKirdy Royal College of Physicians and Surgeons of Glasgow Dr M Jones Royal College of Physicians of Edinburgh Dr A McCune Royal College of Physicians of London Dr M Ostermann Royal College of Physicians of London Dr M Cusack Royal College of Physicians of London Dr T Sabharwal Royal College of Radiologists Mr W Tennant Royal College of Surgeons of Edinburgh Mr J Abercrombie Royal College of Surgeons of England Mr M Bircher Royal College of Surgeons of England

Observers Dr R Hunter Coroners’ Society of England and Wales Mrs J Mooney Healthcare Quality Improvement Partnership Ms T Strack Healthcare Quality Improvement Partnership

118 Trustees Members of the Clinical Outcome Review Programme Mr B Leigh - Chair into Medical and Surgical Care Independent Advisory Dr D Mason - Honorary Treasurer Group: Professor L Regan Dr K Stewart Professor R Endacott Ms R Binks Mr I Martin Professor M Dent Professor T Hendra Dr K Gully Ms J Barber Mrs M Hughes Mr P Lamont Company Secretary - Dr M Mason Professor D O’Donoghue Ms J Russell NCEPOD is a company, limited by guarantee (Company Professor R Taylor number: 3019382) and a registered charity (Charity number: Dr W Taylor 1075588) Mr P Willan Professor K Willett Clinical Co-ordinators Dr I Woods The Steering Group appoint a Lead Clinical Co-ordinator for Dr P Woods a defined tenure. In addition there are 13 Clinical/Nursing Dr M Ferreira Co-ordinators who work on each study. All Co-ordinators Mr T O’Kelly are engaged in active academic/clinical practice (in the NHS) during their term of office. The organisations that provided additional funding to cover the cost of this study: Lead Clinical Co-ordinator Dr M Juniper (Medicine) Aspen Healthcare Clinical Co-ordinators Dr K Wilkinson (Anaesthesia) Beneden Hospital Dr A P L Goodwin (Anaesthesia) BMI Healthcare Mr M Sinclair (Surgery) BUPA Cromwell Mr D O’Reilly (Surgery) East Kent Medical Services Ltd Dr V Srivastava (Medicine) Fairfield Independent Hospital Dr S McPherson (Radiology) HCA International Dr K Horridge (Paediatrics) Hospital of St John and St Elizabeth Dr S Cross (Liaison Psychiatry) King Edward VII’s Hospital Sister Agnes Dr M Allsopp (Adolescent New Victoria Hospital Psychiatry) Nuffield Health Gemma Ellis (Nursing) Ramsay Health Care UK Dr A Michalski (Paediatric Spire Health Care Oncology) St Joseph’s Hospital Supporting organisations The Horder Centre The Clinical Outcome Review Programme into Medical and The London Clinic Surgical Care is commissioned by the Healthcare Quality Ulster Independent Clinic Improvement Partnership (HQIP) on behalf of NHS England, NHS Wales, the Northern Ireland Department of Health, Social Services and Public Safety (DHSSPS), the States of Jersey, Guernsey, and the Isle of Man.

119 Appendices

Appendix 4 – Participation

Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent Abertawe Bro Morgannwg University 5 5 9 7 8 Health Board Aintree Hospitals NHS Foundation Trust 2 1 5 4 4 Airedale NHS Foundation Trust 1 1 3 3 3 Aneurin Bevan University Health Board 2 0 6 5 5 Anglian Community Enterprise (ACE) CIC 0 0 0 0 0 Ashford & St Peter's Hospital NHS Trust 2 2 5 5 5 Aspen Healthcare 4 4 0 0 0 Barking, Havering & Redbridge University 2 2 7 7 7 Hospitals NHS Trust Barnsley Hospital NHS Foundation Trust 1 1 3 3 3 Barts Health NHS Trust 6 6 13 2 1 Basildon & Thurrock University Hospitals 2 2 4 3 3 NHS Foundation Trust Bedford Hospital NHS Trust 1 0 5 4 3 Belfast Health and Social Care Trust 3 2 11 3 7 Benenden Hospital 1 1 0 0 0 Berkshire Healthcare NHS Foundation Trust 5 5 0 0 0 Betsi Cadwaladr University Local Health 19 19 14 5 3 Board Birmingham Community Healthcare NHS 2 2 0 0 0 Trust Blackpool Teaching Hospitals NHS 2 2 5 5 5 Foundation Trust BMI Healthcare 45 41 0 0 0 Bradford Teaching Hospitals NHS 4 4 4 2 1 Foundation Trust Braintree Clinical Services Limited (Serco 0 0 0 0 0 Health Ltd) Bridgewater Community Healthcare NHS 1 1 0 0 0 Trust Brighton and Sussex University Hospitals 3 3 9 9 9 NHS Trust Buckinghamshire Healthcare NHS Trust 6 6 5 3 3 BUPA Cromwell Hospital 0 0 0 0 0 Burton Hospitals NHS Foundation Trust 3 3 4 4 4 Calderdale & Huddersfield NHS Foundation 2 2 10 8 10 Trust Caldew Hospital 0 0 0 0 0 Cambridge University Hospitals NHS 1 1 4 3 1 Foundation Trust

120 Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent Cambridgeshire Community Services NHS 1 0 0 0 0 Trust Cardiff and Vale University Health Board 6 6 5 3 5 Central and North West London NHS 0 0 0 0 0 Foundation Trust Central London Community Healthcare 0 0 0 0 0 NHS Trust Central Manchester University Hospitals 3 3 5 2 1 NHS Foundation Trust Chelsea & Westminster Healthcare NHS 1 0 5 4 2 Trust Chesterfield Royal Hospital NHS Foundation 1 1 3 3 3 Trust City Hospitals Sunderland NHS Foundation 2 2 5 5 5 Trust Colchester Hospital University NHS 1 1 5 5 3 Foundation Trust Countess of Chester Hospital NHS 2 2 4 3 3 Foundation Trust County Durham and Darlington NHS 8 8 8 8 8 Foundation Trust Croydon Health Services NHS Trust 1 1 3 3 3 Cumbria Partnership NHS Foundation Trust 9 9 0 0 0 Cwm Taf Local Health Board 2 2 9 9 9 Dartford & Gravesham NHS Trust 1 1 5 3 3 Derby Teaching Hospitals NHS Foundation 1 1 4 4 4 Trust Derbyshire Community Health Services NHS 11 11 0 0 0 Foundation Trust Doncaster and Bassetlaw Hospitals NHS 2 2 8 5 4 Foundation Trust Dorset County Hospital NHS Foundation 1 1 5 5 5 Trust Dorset Healthcare University NHS 9 9 0 0 0 Foundation Trust East & North Hertfordshire NHS Trust 3 3 6 6 6 East Cheshire NHS Trust 2 2 4 4 4 East Coast Community Healthcare CIC 4 4 0 0 0 East Kent Hospitals University NHS 3 3 12 11 12 Foundation Trust East Kent Medical Services 1 1 0 0 0 East Lancashire Hospitals NHS Trust 5 5 5 4 4

121 Appendices

Appendix 4 – Participation (continued)

Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent East Sussex Healthcare NHS Trust 7 7 7 7 7 Epsom and St Helier University Hospitals 3 3 8 7 6 NHS Trust Fairfield Independent Hospital 1 1 0 0 0 Frimley Health NHS Foundation Trust 3 3 9 8 8 Gateshead Health NHS Foundation Trust 1 1 5 3 2 George Eliot Hospital NHS Trust 1 1 4 4 4 Gloucestershire Care Services NHS Trust 8 8 0 0 0 Gloucestershire Hospitals NHS Foundation 2 2 9 5 5 Trust Great Western Hospitals NHS Foundation 4 4 4 3 3 Trust Guy's & St Thomas' NHS Foundation Trust 2 2 7 7 7 Hampshire Hospitals NHS Foundation Trust 2 0 9 5 2 Harrogate and District NHS Foundation 3 3 5 5 5 Trust HCA International 4 4 1 1 1 Health and Social Services Department, 1 1 2 1 1 States of Guernsey Heart of England NHS Foundation Trust 3 2 15 14 10 Hertfordshire Community NHS Trust 3 3 0 0 0 Hillingdon Hospitals NHS Foundation Trust 1 1 4 2 2 (The) Hinchingbrooke Health Care NHS Trust 1 1 5 5 5 Homerton University Hospital NHS 1 1 3 2 2 Foundation Trust Hospital of St John and St Elizabeth 1 1 1 1 1 Hounslow and Richmond Community 0 0 0 0 0 Healthcare NHS Trust Hull and East Yorkshire Hospitals NHS Trust 2 2 9 9 9 Humber NHS Foundation Trust 0 0 0 0 0 Hywel Dda Local Health Board 9 9 13 11 12 Imperial College Healthcare NHS Trust 4 4 12 9 9 Ipswich Hospital NHS Trust 1 1 4 4 4 Isle of Wight NHS Trust 1 1 3 2 1 Isle of Man Department of Health & Social 1 1 5 5 5 Security James Paget University Hospitals NHS 1 1 2 2 2 Foundation Trust Kent & Medway NHS & Social Care 0 0 0 0 0 Partnership Trust

122 Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent Kent Community Health NHS Trust 10 10 0 0 0 Kettering General Hospital NHS Foundation 1 1 4 4 4 Trust King Edward VII's Hospital Sister Agnes 1 0 0 0 0 King's College Hospital NHS Foundation 3 3 10 10 9 Trust Kingston Hospital NHS Trust 1 1 4 4 4 Lancashire Care NHS Foundation Trust 0 0 0 0 0 Lancashire Teaching Hospitals NHS 2 0 5 4 1 Foundation Trust Lewisham and Greenwich NHS Trust 2 2 9 7 9 Lincolnshire Community Health Services 0 0 0 0 0 NHS Trust Liverpool Heart and Chest Hospital NHS 1 1 4 4 4 Trust Liverpool Women's NHS Foundation Trust 1 1 0 0 0 Locala Community Partnerships CIC 0 0 0 0 0 London North West Healthcare NHS Trust 4 4 10 8 10 Luton and Dunstable Hospital NHS 1 1 5 2 2 Foundation Trust Maidstone and Tunbridge Wells NHS Trust 2 2 6 6 6 McIndoe Surgical Centre 0 0 0 0 0 Medway NHS Foundation Trust 1 1 4 4 4 Mid Cheshire Hospitals NHS Foundation 1 1 5 5 5 Trust Mid Essex Hospitals NHS Trust 1 1 4 3 3 Mid Yorkshire Hospitals NHS Trust 3 3 9 8 9 Milton Keynes Hospital NHS Foundation 1 1 5 4 5 Trust Moorfields Eye Hospital NHS Foundation 1 1 0 0 0 Trust New Victoria Hospital 1 1 0 0 0 Newcastle upon Tyne Hospitals NHS 5 5 7 6 7 Foundation Trust NHS Liverpool Community Health 0 0 0 0 0 Norfolk & Norwich University Hospital NHS 1 1 5 4 5 Trust Norfolk Community Health & Care NHS 7 7 0 0 0 Trust North Bristol NHS Trust 1 1 5 4 5

123 Appendices

Appendix 4 – Participation (continued)

Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent North Cumbria University Hospitals NHS 2 2 10 9 9 Trust North Middlesex University Hospital NHS 1 1 5 5 5 Trust North Somerset Community Partnership 0 0 0 0 0 CIC North Tees and Hartlepool NHS Foundation 2 2 5 4 1 Trust Northampton General Hospital NHS Trust 1 1 4 4 4 Northamptonshire Healthcare NHS 0 0 0 0 0 Foundation Trust Northern Devon Healthcare NHS Trust 15 15 5 4 4 Northern Health & Social Care Trust 3 1 5 3 2 Northern Lincolnshire & Goole NHS 2 2 9 9 9 Foundation Trust Northumbria Healthcare NHS Foundation 8 8 9 7 6 Trust Nottingham University Hospitals NHS Trust 2 2 8 8 8 Nuffield Health 2 2 0 0 0 Outer North East London Community 0 0 0 0 0 Services Oxford Health NHS Foundation Trust 0 0 0 0 0 Oxford University Hospitals NHS Trust 3 0 9 2 2 Oxleas NHS Foundation Trust 0 0 0 0 0 Papworth Hospital NHS Foundation Trust 1 1 2 2 2 Peninsula Community Health CIC 7 7 0 0 0 Pennine Acute Hospitals NHS Trust (The) 4 4 13 13 13 Peterborough & Stamford Hospitals NHS 2 2 5 5 5 Foundation Trust Phoenix Hospital Group 1 1 0 0 0 Plymouth Hospitals NHS Trust 1 1 5 5 5 Poole Hospital NHS Foundation Trust 1 1 0 0 0 Portsmouth Hospitals NHS Trust 1 1 5 4 1 Powys Teaching Local Health Board 9 9 0 0 0 Provide UK 0 0 0 0 0 Queen Victoria Hospital NHS Foundation 1 1 0 0 0 Trust Ramsay Health Care UK 18 12 0 0 0 Robert Jones and Agnes Hunt Orthopaedic 1 1 0 0 0 Hospital NHS Foundation Trust

124 Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent Rotherham Doncaster and South Humber 0 0 0 0 0 NHS Foundation Trust Royal Berkshire NHS Foundation Trust 1 1 4 4 4 Royal Bolton Hospital NHS Foundation Trust 1 1 5 5 5 Royal Bournemouth and Christchurch 1 1 4 4 4 Hospitals NHS Trust Royal Brompton and Harefield NHS 2 2 1 1 1 Foundation Trust Royal Cornwall Hospitals NHS Trust 3 3 5 5 5 Royal Devon and Exeter NHS Foundation 1 1 4 4 4 Trust Royal Free London NHS Foundation Trust 3 3 7 7 7 Royal Liverpool & Broadgreen University 1 1 5 5 5 Hospitals NHS Trust Royal Marsden NHS Foundation Trust (The) 2 2 4 1 1 Royal National Orthopaedic Hospital NHS 1 1 0 0 0 Trust Royal Orthopaedic Hospital NHS 1 1 0 0 0 Foundation Trust Royal Surrey County Hospital NHS Trust 1 1 4 4 4 Royal United Hospitals Bath NHS 2 2 5 5 5 Foundation Trust Salford Royal Hospitals NHS Foundation 1 1 5 3 5 Trust Salisbury NHS Foundation Trust 1 1 5 5 5 Sandwell and West Birmingham Hospitals 2 2 7 6 6 NHS Trust SEQOL(Care and Support Partnership 1 1 0 0 0 Community Interest Company Sheffield Teaching Hospitals NHS 3 3 9 9 8 Foundation Trust Sherwood Forest Hospitals NHS Foundation 2 2 5 5 5 Trust Shrewsbury and Telford Hospitals NHS Trust 2 1 10 9 10 Shropshire Community Health NHS Trust 0 0 0 0 0 Sirona Care & Health CIC 0 0 0 0 0 Solent NHS Trust 2 2 0 0 0 South Devon Healthcare NHS Foundation 1 1 5 5 5 Trust South Eastern Health & Social Care Trust 3 3 4 3 2 South Essex Partnership University NHS 0 0 0 0 0 Foundation Trust

125 Appendices

Appendix 4 – Participation (continued)

Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent South Tees Hospitals NHS Foundation Trust 8 8 4 3 2 South Tyneside NHS Foundation Trust 2 2 5 3 2 South Warwickshire NHS Foundation Trust 1 1 4 4 4 South West London and St Georges Mental 1 1 0 0 0 Health NHS Trust South West Yorkshire Partnership NHS 0 0 0 0 0 Foundation Trust Southend University Hospital NHS 1 1 5 4 1 Foundation Trust Southern Health & Social Care Trust 2 2 10 10 10 Southern Health NHS Foundation Trust 6 6 0 0 0 Southport and Ormskirk Hospitals NHS 1 1 3 1 3 Trust Spire Healthcare 29 14 1 1 1 St George's University Hospitals NHS 1 1 5 5 4 Foundation Trust St Helens and Knowsley Teaching Hospitals 2 2 4 4 4 NHS Trust St Joseph's Hospital 0 0 0 0 0 Staffordshire & Stoke on Trent Partnership 5 5 0 0 0 NHS Trust States of Jersey Health & Social Services 1 1 0 0 0 Stockport NHS Foundation Trust 1 1 3 3 1 Suffolk Community Healthcare 0 0 0 0 0 Surrey & Sussex Healthcare NHS Trust 1 0 5 2 1 Surrey Community Health 0 0 0 0 0 Sussex Community NHS Trust 7 7 0 0 0 Tameside Hospital NHS Foundation Trust 1 1 5 5 5 Taunton & Somerset NHS Foundation Trust 1 1 5 5 5 The Christie NHS Foundation Trust 1 1 0 0 0 The Dudley Group NHS Foundation Trust 1 1 5 5 3 The Foscote Private Hospital 0 0 0 0 0 The Horder Centre 0 0 0 0 0 The Hospital Management Trust 1 1 0 0 0 The Leeds Teaching Hospitals NHS Trust 3 3 9 7 5 The London Clinic 1 1 3 3 0 The Princess Alexandra Hospital NHS Trust 1 1 5 5 5 The Queen Elizabeth Hospital King's Lynn 1 1 5 4 3 NHS Foundation Trust

126 Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent The Rotherham NHS Foundation Trust 1 1 5 5 5 The Royal Wolverhampton Hospitals NHS 2 2 5 5 4 Trust The University Hospitals of the North 2 2 10 6 10 Midlands NHS Trust The Walton Centre NHS Foundation Trust 1 1 2 2 2 Torbay and Southern Devon Health & Care 5 5 0 0 0 NHS Trust Ulster Independent Clinic 1 1 0 0 0 United Lincolnshire Hospitals NHS Trust 3 2 15 9 8 Univ. Hospital of South Manchester NHS 2 2 5 5 5 Foundation Trust University College London Hospitals NHS 4 4 8 5 3 Foundation Trust University Hospital Southampton NHS 1 0 4 4 4 Foundation Trust University Hospitals Birmingham NHS 1 1 5 5 5 Foundation Trust University Hospitals Coventry and 2 2 4 4 4 Warwickshire NHS Trust University Hospitals of Bristol NHS 4 4 5 2 1 Foundation Trust University Hospitals of Leicester NHS Trust 3 3 10 7 10 University Hospitals of Morecambe Bay 3 3 6 6 6 NHS Trust Velindre NHS Trust 1 1 0 0 0 Walsall Healthcare NHS Trust 1 0 4 4 4 Warrington & Halton Hospitals NHS 2 2 0 0 0 Foundation Trust West Hertfordshire Hospitals NHS Trust 2 2 5 5 5 West Middlesex University Hospital NHS 1 0 5 4 5 Trust West Suffolk NHS Foundation Trust 1 1 3 3 3 Western Health & Social Care Trust 2 2 7 2 2 Western Sussex Hospitals NHS Foundation 2 2 6 6 6 Trust Weston Area Health Trust 1 1 5 1 1 Whittington Health 1 1 3 3 3 Wirral University Teaching Hospital NHS 2 2 4 3 3 Foundation Trust Worcestershire Acute Hospitals NHS Trust 3 3 6 3 6

127 Appendices

Appendix 4 – Participation (continued)

Trust/Health Board name Number of Number of Number of Number Number of sets participating organisational included of clinician of case notes hospitals questionnaires cases/ clinician questionnaires returned returned questionnaires returned sent Worcestershire Health and Care NHS Trust 0 0 0 0 0 Wrightington, Wigan & Leigh NHS 2 2 4 3 3 Foundation Trust Wye Valley NHS Trust 2 1 5 5 5 Yeovil District Hospital NHS Foundation 1 1 5 4 4 Trust York Teaching Hospitals NHS Foundation 5 5 9 9 9 Trust

128 Published November 2015 by the National Confidential Enquiry into Patient Outcome and Death

Ground Floor Abbey House 74-76 St John Street London EC1M 4DZ

T 0207 251 9060 F 0207 250 0020 E [email protected] w www.ncepod.org.uk

ISBN 978-0-9926184-4-5

A company limited by guarantee Company no. 3019382 Registered charity no. 1075588